Slashdot Mirror

An anonymous reader writes: Discovered in 1997 by Aaron Spangler and never fixed, the WinNT/Win95 Automatic Authentication Vulnerability (IE Bug #4) is certainly an excellent vintage. In Windows 8 and 10, the same bug has now been found to potentially leak the user's Microsoft Live account login and (hashed) password information, which is also used to access OneDrive, Outlook, Office, Mobile, Bing, Xbox Live, MSN and Skype (if used with a Microsoft account). The bug itself seems to be present in all Windows systems since Windows 95 / NT, although only Windows 8 and above are effectively compromised. To see if your machine is affected, you may want to check the public demonstration of the exploit, set up by the guys from [Perfect Privacy] and based on [VladikSS] original work. Basically, the default User Authentification Settings of Edge/Spartan (also Internet Explorer, Outlook) lets the browser connect to local network shares, but erroneously fail to block connections to remote shares. To exploit this, an attacker would simply set up a network share. An embedded image link that points to that network share is then sent to the victim, for example as part of an email or website. As soon as the prepped content is viewed inside a Microsoft product such as Edge/Spartan, Internet Explorer or Outlook, that software will try to connect to that share in order to download the image. Doing so, it will silently send the user's Windows login username in plaintext along with the NTLMv2 hash of the login password to the attacker's network share.

An anonymous reader writes with good news for advocates of Full Disclosure of security vulnerabilities. A week ago, the venerable full-disclosure list was shut down; now, a successor has arisen run by fyodor. From the announcement email: "As an F-D subscriber and occasional poster myself, I was as shocked as you all last week when John Cartwright threw in the towel and shuttered the list. Now I don't blame him one bit. He performed a thankless job admirably for 12 years and deserves some time off. But I, for one, already miss Full Disclosure. So I decided to make a new list today which is a successor in name and spirit. Like the old one, it uses Mailman and is being archived by my Seclists.org site as well as numerous other archives around the world. This list is a fresh start, so the old userbase won't automatically transfer over. And I haven't added any of you either, because it is your choice. ... I hope you'll join us and resume posting your security info and advisories. If not now, then someday."

brothke writes "The 1962 song Wipe Out, with its energetic drum solo started, was the impetus for many people to take up playing the drums. Similarly, Nmap, the legendary network scanner, likely interested many in the art of hacking, and for some, started a career for security professionals and hackers. Nmap and its creator Fyodor need no introduction to anyone on Slashdot. With that, Nmap Network Scanning: The Official Nmap Project Guide to Network Discovery and Security Scanning, is a most useful guide to anyone interested in fully utilizing Nmap." Read on for the rest of Ben's review. Nmap Network Scanning: The Official Nmap Project Guide to Network Discovery and Security Scanning author Gordon Lyon (Fyodor) pages 468 publisher Nmap Project rating 9 reviewer Ben Rothke ISBN 978-0979958717 summary Valuable book about an invaluable security tool One may ask, why spend $50 on this book, when the Nmap Reference Guide provides a significant amount of the basic information needed to use the tool, especially since the reference guide is both free, and well written. The reference guide is included in the book in chapter 15, and takes up 41 pages. And for those that are cash strapped, the free reference guide is the way to go.

In addition, the web site for the book notes that about half of the content is available in the free online edition. The most useful information is in the book in chapters exclusive to the print edition, which includes Detecting and Subverting Firewalls and Intrusion Detection System, Optimizing Nmap Performance, Port Scanning Techniques and Algorithms, Host Discovery, and troubleshooting.

The main benefit of the buying the book is that it has the collected wisdom of Fyodor's, in addition to numerous real-world scenarios, and Nmap commands not documented elsewhere. At over 400 pages, the books 15 chapters provide the reader with everything they need to know about using Nmap to the fullest.

Chapter 1 starts with an overview of the history of Nmap and how it came to be. As to the question of whether port scanning is legal, the author writes that it is best to avoid the debate and its associated analogies. He advises that it's best to avoid ISP abuse reports and criminal charges, by not annoying the target network administrators in the first place. Chapter 1 provides a number of practical suggestions on just how to do that.

A complaint against Nmap it that is has often been blamed for crashing systems. Chapter 1 shows that the reality is that Nmap will rarely be the primary cause of a system crash. The truth is that many of the systems that crashed as a result of an Nmap scan were likely unstable from the outset, and Nmap either pushed them over the top or they coincidentally crashed at the same time as the Nmap scan.

An ironic incident detailed in chapter 3 is when someone from the information security department of Target Corp. complained to the author that he felt the Nmap documentation was particularly directed at his organization; given the use of the term target. He requested that the Nmap documentation be changed from targetto example. The section on target enumeration in the book shows the author did not take that request to heart.

Another example of where the book goes beyond what is in the reference guide is where the author shows the most valuable TCP ports via his probe of tens of millions of IP addresses across the internet. Not surprisingly, ports 80 23 and 443 were the top three most commonly open TCP ports. It is surprising that other ports, which should have been secured long ago, are still as vulnerable as ever.

For the serious Nmap user, the book is worth purchasing just for the indispensable information in chapter 16, which is about optimizing Nmap performance. The author writes that one of his highest priorities in the creation of Nmap has been performance. Nmap uses parallelism and numerous advanced algorithms to execute its blazingly fast scans. This chapter shows how to create Nmap commands to obtain only the information you care about and significantly sped up the scan. The chapter details numerous scan time reduction techniques, and strategies on how to deal with long scans. The chapter concludes with the output of a user who, with a customized Nmap command, was able to reduce his scan of a 676,352 IP address network from nearly a week to 46 hours.

Chapter 10 is also a fascinating chapter on the topic of detection and subverting of firewalls and IDS. The function of such tests on an internal network is to help an organization understand the dangers and risks of a real attack. Since it is not uncommon for firewalls to be accidentally misconfigured, or have rule bases that leak from far too many rules; such a test can be quite useful to any network.

Nmap Network Scanning: The Official Nmap Project Guide to Network Discovery and Security Scanning is the guide for anyone who wants to get more out of Nmap. It is useful whether one is a novice and only getting into basic security testing, or an advanced user looking for ways to optimize Nmap.

The book takes a real-world approach on how to use the tool and clearly documents every Nmap feature and option. It also shows how the tool should be correctly used in various settings.

What is unique about is that this is a rare book in which the creator of the program wrote it. Linus Torvalds never got around to writing a Linux reference, nor did the creators of the Check Point firewall. In Nmap Network Scanning, the reader gets the story from the creator of the code itself. This then is the ultimate Nmap reference guide.

Aside from the history and use of the program in the first chapter, the rest of the book is an extreme guide to maximizing the use of Nmap. It is written by a programmer and written for the technically astute. Anyone who wants to maximize their use of Nmap will find no better reference.

Ben Rothke manages the Bright Hub Enterprise Security channel and is the author of Computer Security: 20 Things Every Employee Should Know.

You can purchase Nmap Network Scanning: The Official Nmap Project Guide to Network Discovery and Security Scanning from amazon.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.

Cuts and bruises writes "Hacker Joanna Rutkowska has flagged a "very severe hole" in the design of Windows Vista's User Account Controls (UAC) feature. The issue is that Vista automatically assumes that all setup programs (application installers) should be run with administrator privileges — and gives the user no option to let them run without elevated privileges. This means that a freeware Tetris installer would be allowed to load kernel drivers. Microsoft's Mark Russinovich acknowledges the risk factor but says it was a 'design choice' to balance security with ease of use."

TheViewFromTheGround writes "Fyodor of nmap fame has released a top 100 list of network security tools, based on a poll of the nmap-hackers list, each with a handy synopsis and useful information about source-code availablity and OS-compatibility. The last version of this survey was published in 2003."

TheViewFromTheGround writes "Fyodor of nmap fame has released a top 100 list of network security tools, based on a poll of the nmap-hackers list, each with a handy synopsis and useful information about source-code availablity and OS-compatibility. The last version of this survey was published in 2003."

NoExec writes "Hot off the nmap-hackers list comes news that the Nmap Security Scanner version 4.00 has been released. Dozens of major changes since 3.50 (2 years ago) are listed in the announcement. These include a rewritten (for speed and memory efficiency) port scanning engine, ARP scanning, a brand new man page and install guide, 'l33t ASCII art, runtime interaction, massive version detection improvements, MAC address spoofing, increased Windows performance, 500 new OS detection fingerprints, completion time estimates, and much more."

NoExec writes "Hot off the nmap-hackers list comes news that the Nmap Security Scanner version 4.00 has been released. Dozens of major changes since 3.50 (2 years ago) are listed in the announcement. These include a rewritten (for speed and memory efficiency) port scanning engine, ARP scanning, a brand new man page and install guide, 'l33t ASCII art, runtime interaction, massive version detection improvements, MAC address spoofing, increased Windows performance, 500 new OS detection fingerprints, completion time estimates, and much more."

NoExec writes "Hot off the nmap-hackers list comes news that the Nmap Security Scanner version 4.00 has been released. Dozens of major changes since 3.50 (2 years ago) are listed in the announcement. These include a rewritten (for speed and memory efficiency) port scanning engine, ARP scanning, a brand new man page and install guide, 'l33t ASCII art, runtime interaction, massive version detection improvements, MAC address spoofing, increased Windows performance, 500 new OS detection fingerprints, completion time estimates, and much more."

spafbnerf writes "Fyodor, author of the open-source network scanning tool Nmap, posted a story to the nmap-hackers list about having received a number of subpoenas from the FBI this year, demanding webserver log data, none of which produced anything, either because they sought old information that had already been deleted from his logs, or because the subpoenas were improperly served. In every case the request was narrowly crafted, usually directed at finding out who visited the site in a very short window of time, such as a five minute period. Fyodor writes: "If they see that an attacker ran the command "wget http://download.insecure.org/nmap/dist/nmap-3.77.t gz" from a compromised host, they assume that she might have obtained that URL by visiting the Nmap download page from her home computer"." Update: 11/25 20:21 GMT by T : Reader kv9 adds a link to Kevin Poulson's story at SecurityFocus.

spafbnerf writes "Fyodor, author of the open-source network scanning tool Nmap, posted a story to the nmap-hackers list about having received a number of subpoenas from the FBI this year, demanding webserver log data, none of which produced anything, either because they sought old information that had already been deleted from his logs, or because the subpoenas were improperly served. In every case the request was narrowly crafted, usually directed at finding out who visited the site in a very short window of time, such as a five minute period. Fyodor writes: "If they see that an attacker ran the command "wget http://download.insecure.org/nmap/dist/nmap-3.77.t gz" from a compromised host, they assume that she might have obtained that URL by visiting the Nmap download page from her home computer"." Update: 11/25 20:21 GMT by T : Reader kv9 adds a link to Kevin Poulson's story at SecurityFocus.

acz writes "The brain and guts driving the development of Kismet is Mike Kershaw alias Dragorn, who works during the day on IBM mainframes and hacks code at night. Kismet is simply the best war driving tool out there plus it's free as in GPL and can even run on your linux PDA. In a recent interview posted on HERT today, he says: 'I've become entirely jaded towards security as a whole (or rather, people's complete lack of it) and not much surprises me when it comes to open wireless networks. ... the overall percentage of unencrypted networks is still at about 80%.'"

Kinetic writes "There are many great open source security tools out there, Nmap, Nessus, and DSniff, just to name a few. However, with the world of security constantly changing, this begs the question, what open source security tools are missing? What commercial security tools have no viable open source alternatives? When securing/testing/exploring networks (home or enterprise), what security tools/applications/functionality are lacking (or non-existent) in the open source world?"

reddigitaldragon asks: "Some people re-install once a year, but if you're anything like me your machine is formatted at least once a month. After the OS is in, then come the favorite/must have/most used programs to install. My first installations for Windows (I use it; get over it): Trillian, Winrar, Firefox, Winamp, SmartFTP, Azureus, NMap, GKrellM, PowerDVD. What are your First 10 installed programs?" What are the first 10 programs you would install on a Windows machine? How about for a Unix machine?

ronys writes "The venerable USENIX organization has written a fine response to SCO's letter to Congress. As they point out, 'USENIX was here before SCO. USENIX was here before Linux.' Short and well written." And Reece Arnott writes: "As part of the NMap Press Release for the latest version of NMap, is a statement that explicitly revokes SCO's licence to redistribute it. From the press release: 'SCO Corporation of Lindon, Utah (formerly Caldera) has lately taken to an extortion campaign of demanding license fees from Linux users for code that they themselves knowingly distributed under the terms of the GNU GPL. They have also refused to accept the GPL, claiming that some preposterous theory of theirs makes it invalid (and even unconstitutional)! Meanwhile they have distributed GPL-licensed Nmap in (at least) their "Supplemental Open Source CD". In response to these blatant violations, and in accordance with section 4 of the GPL, we hereby terminate SCO's rights to redistribute any versions of Nmap in any of their products, including (without limitation) OpenLinux, Skunkware, OpenServer, and UNIXWare. We have also stopped supporting the OpenServer and UNIXWare platforms.'"

Anonymous Coward writes "Up until now, everyone's favorite port scanner, nmap has had decent OS detection (through TCP fingerprinting) and service identification based on the open port, but the latest version, 3.45 released today, has version detection for each service! This means not only can nmap tell you that httpd is running on port 80, but that it is `apache httpd version 2.0.39`! While this is a little bit worrisome because of what malicious purposes people might use nmap's version detection for, this should make the jobs of admins everywhere easier and keep us all more on our toes when it comes to security. Fyodor has also published a paper on how the version detection works."

Anonymous Coward writes "Up until now, everyone's favorite port scanner, nmap has had decent OS detection (through TCP fingerprinting) and service identification based on the open port, but the latest version, 3.45 released today, has version detection for each service! This means not only can nmap tell you that httpd is running on port 80, but that it is `apache httpd version 2.0.39`! While this is a little bit worrisome because of what malicious purposes people might use nmap's version detection for, this should make the jobs of admins everywhere easier and keep us all more on our toes when it comes to security. Fyodor has also published a paper on how the version detection works."

Anonymous Coward writes "Up until now, everyone's favorite port scanner, nmap has had decent OS detection (through TCP fingerprinting) and service identification based on the open port, but the latest version, 3.45 released today, has version detection for each service! This means not only can nmap tell you that httpd is running on port 80, but that it is `apache httpd version 2.0.39`! While this is a little bit worrisome because of what malicious purposes people might use nmap's version detection for, this should make the jobs of admins everywhere easier and keep us all more on our toes when it comes to security. Fyodor has also published a paper on how the version detection works."

Eric Stats writes "Working as a Network Engineer for web-hosting company that prides itself on uptime and network availability, and moonlighting as a part-time Linux administrator, my managers and clients are starting to expect a level of information security knowledge from me. I decided that if I wanted to take my career to the next level, I needed to develop some security-specific skills. I heard a lot about the open source Intrusion Detection System (IDS), Snort from friends and co-workers (mostly that it was a pain to get running, and an even bigger pain to understand what it was doing)." To get past those frustrations, Eric looked at two more books on Snort (and compares them to the already-reviewed Intrusion Detection with Snort ); read on below for his take on what each offers. Intrusion Detection with SNORT: Advanced IDS Techniques Using SNORT, Apache, MySQL, PHP, and ACID; Intrusion Detection with Snort; Snort 2.0 Intrusion Detection author (See each) pages (See each) publisher (See each) rating (See each) reviewer Eric Stats ISBN (See each) summary (See each)

I ran Snort at home for a while, using the online docs, but I could never get a handle on which output plugin to use (When to log? When to alert?), how to email alerts to myself (I later found out Snort doesn't natively do this), and how to create signatures from packet captures (no online docs at all for this). When I did get The Pig running, it filled up my log directory with thousands of small alert files, which ended up being in tcpdump format. This frustrated the hell out of me, so I decided I needed to find a good book on Snort, as the online docs simply did not describe how to use Snort from start to finish.

In the past few months, an assortment of books have come out on Snort. Because it has begun to eclipse closed-source, multimillion dollar IDSes in terms of raw performance and features, much attention is currently focused on Snort. Naturally, when an open source project achieves this level of notoriety, publishers, venture capitalists, and corporations want to get in on the game. The flood of Snort books is a testament to this, but it doesn't mean they were all created equally. This book review covers the three books on Snort currently available (we will see another two Snort books later this winter). It covers what is good about them, what is bad, and who the target audience is for each. If you are looking to learn intrusion detection the open source way, or simply do not have a million-dollar IT security budget, these books are a good starting point.

Each of these three books serves a different purpose and consequently is appropriate for a different reader. In summary, Rafeeq Rehman's Intrusion Detection with Snort: Advanced IDS Techniques Using SNORT, Apache, MySQL, PHP, and ACID presents a concise, quick-start guidebook to getting Snort up and running fast. He doesn't delve into the details of Snort, and this book makes a perfect choice for a reader who wants to get The Pig up and running quickly and move on to something else.

The whole gaggle of authors that put together Snort 2.0 Intrusion Detection created a much-needed user manual for Snort. This book makes for good desktop reference, but assumes you understand the core concepts of intrusion detection, or have significant field experience with Snort. It is also somewhat convoluted to read; I suppose it's inevitable when you have 12 authors working on a single book, it is going to come out somewhat disjointed and jumbled. If I hadn't read the other two books first, I doubt I would have been able to piece together what this book is talking about in places. (Such as referring to Barnyard logs in one chapter and "unified binary format" in another; how is the reader going to know they are the same?)

Lastly, Jack Koziol's Intrusion Detection with Snort is a guidebook for using Snort in the real world, either on small networks or in large corporate settings. Like any security tool, Snort is only as effective as its operator. Snort can do an enormous number of things, but if you don't understand the "how and why" you aren't going to be able to apply your knowledge in unexpected, different, or new situations. Koziol's book bridges the gap and teaches you the nitty-gritty Snort details not found in online docs, as well as how to apply your newfound IDS knowledge in practice. This book does lack in terms of screenshots and diagrams, which can be frustrating at points. Instead of a paragraph of text, a simple diagram would have sufficed.

Intrusion Detection with SNORT: Advanced IDS Techniques Using SNORT, Apache, MySQL, PHP, and ACID author Rafeeq Rehman pages 288 publisher Prentice Hall rating 7/10 ISBN 0131407333

I first picked up Rehman's Intrusion Detection with Snort: Advanced IDS Techniques Using SNORT, Apache, MySQL, PHP, and ACID. Rehman's book is also a member of the Bruce Perens Open Source Series. All of the books in his series are published under the OPL. Overall, Rehman's book served as a good intro to Snort. I followed the examples, used some of the custom startup and log-rotation scripts, and got Snort working for the first time. I also learned of ACID, which is a PHP-based GUI for Snort, put out by Carnegie Mellon's CERT/CC. It makes managing alerts from Snort much less time-intensive. It was an exciting experience, but the book left me in the dark on a number of concepts that I knew I needed to learn. I still didn't understand what I was getting out of Snort; I had so many alerts I couldn't "tune out the noise." I didn't know when to use log or alert plugins, so I just turned on both for safety's sake. I also found that Snort was dropping packets (meaning it wasn't able to keep up with the traffic load going to my webservers hosted at home), but didn't find any way to fix this problem. This setup was fine for experimenting at home, but I didn't feel I would be able to use Snort in a mission-critical corporate setting yet.

Intrusion Detection with Snort author Jack Koziol pages 400 publisher SAMS Publishing rating 9/10 ISBN 157870281X

I thumbed through Jack Koziol's Intrusion Detection with Snort at the bookstore, and it seemed to have some more detailed descriptions of using Snort. It also had a lot of the planning, deployment, and maintenance activities you never think of until you are faced with one at 2 a.m. (such as how to upgrade Snort in an organized manner after a vicious integer overflow exploit is released for a core Snort component). It is also the most popular Snort book, so I figured I would buy it. When I took it home, I learned where to place Snort on a network, and what advantages and disadvantages there are to different IDS sensor placement strategies, something I had never considered.

Koziol's book also had the technical detail I was in desperate need of. I learned how to use Barnyard to spool alerts, which keeps Snort from dropping packets. I got to write my own attack signatures from scratch by using Ethereal packet captures in an controlled lab environment. I created a targeted ruleset; it enables specific attack signatures based on what I actually have running on my network, simply using nmap and some complicated perl scripts. The targeted ruleset went a long way to reducing false alerts, and is now a selling product from the Snort commercial vendor, Sourcefire. I finally got email alerts working using syslog-ng with Snort. The book ends with some more advanced content, namely using Snort as an Intrusion Prevention device. You can setup Snort to block packets that match a signature, using Inline Snort, or you can have Snort reconfigure routers and firewalls to block offending IP addresses, using SnortSam. I've experimented with Inline Snort as part of a honeypot, but, as the author points out, this is not yet production-safe, as it can easily be used by attackers to disrupt network availability.

Snort 2.0 Intrusion Detection authors Jay Beale, Anne Carasik, Aidan Carty, Scott Dentler, Adam M. Doxtater, Wally Eaton, Jeremy Faircloth, James C. Foster, Vitaly Osipov, Jeffrey Posluns, Ryan Russell, Brian Caswell pages 485 publisher Syngress rating 4/10 ISBN 1931836744

The final Snort book in this review is Snort 2.0 Intrusion Detection. This book has a lot of the screenshots and figures that the Koziol and Rehman books leaves out. It also contains a lot of useful diagrams, about one for every other page, and a CD-ROM with all of the Snort source and a pdf version of the book. This book, and the Koziol book, cover Snort version 2.0, which isn't all that much different from version 1.9 covered in the Rehman book. Still, it is nice to have the most up-to-date documentation, but it doesn't make the Rehman book any less effective. This book has the most reference material in it, over 500 pages' worth, and it has very organized user manual-like descriptions of important Snort components (preprocessors, output plugins, and rules). Keep in mind that this book was created more as a user manual rather than an implementer's guide. You aren't going to see planning, deployment, and maintenance activities as well as technical deployment examples, as in the Koziol book. And, you aren't going to find a concise quick-start guide such as the Rehman book.

In summary, you aren't going to find anything in this book that isn't in the other two. What you will find is lengthy descriptions, and a lot more screenshots. As stated before, Snort 2.0 Intrusion Detection was written by 12 different people (one of them a Sourcefire employee and Snort.org website maintainer, Brian Caswell). This is obviously done by the publisher to get the book out as fast as possible, which is important for technology book publishers as books are outdated quickly, but has the end result of a disjointed book that contradicts itself in many areas. An example: one author stresses how deadly important it is for us to only use the latest Snort version, while another tells us to use the CDROM that comes with the book, which contains an outdated version of Snort.

You can clearly tell a different authors worked on different chapters, as the style and format change frequently. You can also tell that the authors didn't talk to each other much, as you will find one author referring to something in one chapter (unified binary format) that he expected to have been explained in a previous chapter. In print, the concept was not explained until later, which can be really frustrating if you are not a Snort pro. Additionally, there are enough grammatical errors in the book to be distracting, and, much like a vendor-provided user manual, the chapters don't logically flow from one to the next. If you do purchase this book, this slashdotter would recommend it as a supplement to either the Rehman or Koziol book.

You can purchase Intrusion Detection with SNORT: Advanced IDS Techniques Using SNORT, Apache, MySQL, PHP, and ACID , Intrusion Detection with Snort , and Snort 2.0 Intrusion Detection from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.

Slashback tonight with more on model rocketry (and metaphysical rocketry to boot), Metallica's music online -- this time voluntarily, the fall of Ars Digita, nmap's reaction to SCO, and more. Read on for the details.

How is this sanguine? peterb writes "Slashdot has previously reported on Eve Andersson's whitewash of Ars Digita. Her screed placed responsibility for all the problems fully on the shoulders of the Venture Capitalists, while ignoring the role of those that asked the VCs for money. Ars Digita's Michael Yoon has a somewhat more sanguine and less hysterical version of the same story."

I wonder if shoulder chips can be recycled as fuel ...

All them perls don't come cheap. dogma01 writes "It's been almost a year since I submitted this story on Slashdot about the Perl Foundation Fund Drive. With a new year there has been a new round of grants. Every dime helps improve the community and bring us one step closer to Perl6. Please donate here."

The largest model is actually the one that's currently in use. joshamania writes "I knew when I saw the first post about the 'largest scale model of the solar system' I should have piped up. The second post has driven me over the edge. I call shenanigans! The Maine model is not the largest, and Peoria, IL, my hometown, has had the largest model for many years now, the Pluto model (in Kewanee, IL) being over 60 miles away from the sun model. In fact, a bicycle tour of the model is organized every summer and reoccurs in August."

Still at maximum. Danta writes "As the QNX site seems to have received an indirect slashdotting, here is a BitTorrent link to the free version of the QNX OS."

And what's in your makefile? JediTrainer writes "Community backlash begins! The author of Nmap has decided to remove all support of the SCO operating system as of version 3.28. Quoting the changelog, 'SCO operating systems are no longer supported due to their recent (and absurd) attacks against Linux and IBM. Bug reports relating to UnixWare will be ignored, or possibly even laughed at derisively. Note that I have no reason to believe anyone has ever used Nmap on SCO systems. Unixware sucks.'"

Speaking of backlash ... Ransak writes "Speak out! Space-Rockets.com has started a letter/fax campaign to sway political opinion, but needs your help! This hobby enjoyed by thousands of future scientists and astronaunts has been put in serious jeopardy by bad legislation. Senator Herb Kohl was one of the coauthors of the Safe Explosives Act, who not surprisingly, is blocking an amendment to ease restrictions on model rocketry. Wisconson geeks, take note of your Senators actions!"

... and speaking of rockets: BuR4N writes "The x-prize foundation has decided not to accept an application from a Budapest based team called GCT (Gravity Control Technologies) due to their highly questionable proposal. GCT pitched a "propellantless propulsion technology" that quote "is capable of controlling gravity for flight". Here is the full story. It would be very interesting to hear from the scientific community if this is just silliness or something that eventually could lead anywhere.."

I hope these guys don't take up making ham sandwiches. acidblood writes "Following up on yesterday's story concerning ice cream and liquid nitrogen, it appears someone was keen to try it out, and this is the result."

I can't tell from the page when exactly this was made. Whether it was truly in response to Gray's recipe or not, this site certainly provides more amusing visual aids.

The medium is the message, or something like that. LineNoiz writes "There is an interesting article over at MSNBC outlining Metallica's attempt to take advantage of the internet as a music distribution medium. It seems their newest album 'St. Anger' has a code on it which can be used to access their "Audio Vault" where users can download MP3 recordings of live concerts. The site's motto? 'Download. Burn. Share. Kick Ass.' Is this just a flagrant attempt to recapture the interest of the thousands of fans they lost in their battle with Napster, or a genuine good idea?"

Readers may recall this interview with Metallica's Lars Ulrich.

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

You asked nmap creator Fyodor many excellent questions, and his answers (below) are just as excellent. You'll want to set aside significant time to read and digest this interview, because Fyodor didn't just toss off a few words, but put some real time and energy into his answers.

1) Interesting stories involving nmap?
by Neologic

Nmap has obviously become a huge success in the *nix world. I would wager that practically all sysadmins and security folk use nmap. With this sort of use by such creative and lazy people, there must have been some interesting stories involving nmap, perhaps unusual uses of it, or funny anecdotes. Are there any you would like to share?

Fyodor

The coolest use ever was undoubtedly when Trinity used it to try and save the human race :). But the use I find most gratifying are the Chinese students and residents who have written me about how they use Nmap to locate open proxies. These proxies allow for surfing the uncensored Internet, including the news, educational, pornographic, religious, open source software, government, political, search engine, and human rights sites that are blocked by the Great Firewall of China.

Many of the best features in Nmap came from the user community in ideas if not implementation. For example, the protocol scan (-sO) determines what IP protocols (TCP, UDP, GRE, etc.) a host is listening for. I had not thought of this, but the idea and patch came out of the blue one day in an email from Gerhard Rieger. On another day, a guy named Saurik sent a patch called Nmap+V that allows Nmap to do basic service/version fingerprinting against open ports. It has attracted a cult following, and I plan to add similar functionality to Nmap this year. The initial Windows port by eEye arrived similarly. Despite all these great suggestions, certain other user-contributed ideas are not on the agenda.

Then there are a small handful of users who detect problems nobody else would ever notice, like 4 byte/host memory leaks. They send me error messages with notes saying the bug happens "about once per 700,000 IPs". I have no idea what these guys are up to, but some have been sending me this kind of mail for years. They can't be spammers, as they are intelligent and also use more sophisticated scan techniques than you would need to just find SMTP servers.

2) Recent increases in anal-retentiveness...?
by Zeriel

There's been a marked increase in system administrators thinking that anything even remotely resembling a network scan is eeeeevil (case in point, last year I almost got kicked out of college for scanning port 80 on my dorm subnet looking for interesting websites to read)...

What do you think can be done to make scanning IP addresses/ports have less of a negative stigma? This is in the same sort of category as legit vs. illegit uses of anything else (P2P, whatever)--what's the rationale for punishing something that could maybe lead to criminal activity, and how can we make network scanning tools have practical uses again?

Fyodor

That is an excellent question, and one that concerns me as well. But first, I think your final statement is too extreme. I would guess 90% of network scanning is non-controversial. You will rarely be badgered for scanning your own machine or the networks you administer. The controversy comes when scanning other networks. There are a lot of (good and bad) reasons for doing this sort of network exploration. Perhaps you are scanning the other systems in your {dorm, department, cable LAN, conference LAN} to look for publicly shared files (FTP, SMB, WWW, etc.). Or perhaps your just trying to find the IP of a certain printer. Maybe you scanned your favorite web site to see if they are offering any other services, or because you are curious what OS they run. Perhaps you are just trying to test connectivity, or maybe you wanted to do a quick security sanity-check before handing off your credit card details to that ecommerce company. You might be conducting Internet research, or be bored on a rainy afternoon. Or are you conducting reconnaissance in preparation for a breakin attempt?

The remote administrators rarely know your true intentions, and do sometimes get suspicious. The best approach is to get permission first. I've seen a few people with non-administrative roles land in hot water after deciding to "prove" network insecurity by launching an intrusive scan of the entire company or campus. Admins tend to be more cooperative when asked in advance than when woken up at 3AM by an IDS alarm claiming they are under massive attack.

You compared Nmap to P2P tools in having a "negative stigma". In both cases, one effective way to fight the stigma is to limit your own use to "legitimate" purposes. Use BitTorrent to download RedHat ISOs, but not Matrix Reloaded. Use Nmap to secure and monitor your computers, but not to attack other networks. And if you decide to attack other networks anyway, please be courteous and set the evil bit.

Now I'll admit that I don't always obtain explicit permission before scanning other networks. I don't believe (but IANAL) that a simple port/OS scan of a remote system is or should be illegal. Any machine connected to the Internet will be scanned so often that most admins ignore such "white noise" anyhow. But scan other networks often enough, and someone will eventually complain. So my advice would be:

1. Don't do anything controversial from your work or school connections. Even though your intentions may be good, you have too much to lose if someone in power (boss, dean) decides you are a malicious cracker. Do you really want to explain your actions to someone who may not even understand the terms "port scanner" or "packet"? Spend $10 bucks a month for a dialup or shell account. You didn't really violate this rule, as scanning your dorm subnet for just port 80 should not even be remotely controversial!
2. Target your scan as tightly as possible. If you are only looking for web servers, specify -p80 rather than scanning all 65,535 TCP ports on each machine. If you are only trying to find available hosts, do an Nmap ping scan. Don't scan a /16 when a /24 will suffice. The random scan mode now takes an argument specifying the number of hosts, rather than running forever. So consider -iR 1000 rather than -iR 10000 if the former is sufficient. Use the default timing (or even "-T Polite") rather than "-T Insane".
3. Nmap offers many options for stealthy scans, including source-IP spoofing, decoy scanning, and the more recent Idle Scan technique. But remember there is always a trade-off. You will be harder to detect if you launch scans from an open WAP far from your house, with 17 decoys, while doing followup probes through a chain of 9 open proxies. But if anyone (such as Tsutomu Shimomura) does track you down, they will be mighty suspicious of your intentions.

I occasionally consider adding some sort of "notification packet" prior to a scan that would give hosts the chance to respond and opt-out. This would be like the /robots.txt directives currently used to control polite Web robots. Perhaps the format could even include a text string that IDS systems could log, like: nmap -sS -p- -O -m "Direct questions about this scan to ops at x3512" 192.168.0.0/16 nmap -sS -p- -O -m "mY n4m3 iZ Zer0 |<00L and I'll 0wn j0o%#@" targetcompany.com/24 Of course Nmap would have an option to omit the notification or to send it and ignore any negative responses. Some scanners, such as ISS Internet Scanner already send out NetBIOS popup messages to scanned hosts by default, and other scanners use syslog. I won't be adding any features like this to Nmap unless I see substantial demand and the obvious issues are worked out.

3) OS fingerprinting
by neoThoth

What are the latest advances in fingerprinting networked devices that seem most promising to you? I have started reading papers on HTTP fingerprinting and such and wonder how these will figure into the NMAP architecture. What are the most elusive OS's that aren't on the NMAP OS fingerprint database?

Fyodor

There are a number of OS detection techniques I hope to add this year. One is to guess (or calculate) the initial TTL of response packets, since this varies by OS. Some operating systems also "reflect" your own chosen TTL under various circumstances. Then there are some newer TCP options, such as selective ack that I might test for. Explicit Congestion Notification (RFC 2481/3168) also shows promise. I'll probably add all of these at once later this year, after discussions with the Nmap-dev list. If you wish to participate, you can join that list by sending a blank email to nmap-dev-subscribe@insecure.org. There is also a low volume, moderated list for announcements about Nmap, Insecure.org, and related projects. You can join the 15,000 current members by mailing nmap-hackers-subscribe@insecure.org [archives].

While adding new fingerprinting techniques is fun and exciting, improving the signature database often ads more value. The DB now contains more than 850 signatures, from the Acorn RISC OS and Aironet wireless LAN bridge to the ZoomAir wireless gateway and Zyxel Prestige routers. We're talking gaming consoles, phones, PBX systems, PDAs, webcams, networked power switches, you name it! New fingerprints are submitted daily.

Application level fingerprinting (including HTTP) is coming. I usually regret stating dates, but I hope to develop this functionality within the next 3 months.

4) Stepping into a network security career
by Anonymous Coward

I'll be graduating this month with a shiny new BS in Computer Science. I've done plenty of Unix sysadmin work throughout college and even deployed some high-interaction honeynets. I'm very interested in network security and systems programming. Do you have any advice for people in my situation who want to head into a career in network security?

Fyodor

Congratulations on your graduation! Unfortunately (for newcomers), the security field is one that often expects substantial experience and references. This is partly because these jobs require extraordinary trust, and also because of an aversion to mistakes. Everyone makes mistakes, but they can be extraordinarily costly in security and neophytes tend to make more of them. But don't lose hope! Talented security minds are still in very high demand, just be aware that you will have to work even harder to prove yourself.

Here are my suggestions for anyone starting out in network security, whether for fun or profit:

Step 1: Learn everything you can

1. You may wish to start with reading a general overview of security, such as Practical Unix and Internet Security 3rd Edition.
2. Reading alone won't teach you much. Hands-on experience is critical, so I would set up at least a basic test network. At the very minimum you should have a Unix box or two and a Windows machine (because these are very common in the real world). You can use very cheap machines, or even emulate a large network with virtualization software such as VMWare.
3. Next you should learn more about how attacks are performed. Take a look at the excellent and free Open Source Security Testing Methodology Manual (OSSTMM). This document aims to provide a comprehensive framework for security testing. But it mostly lists tasks to perform, without specifying how to do so. You will gain a lot from this manual if you research the tasks you don't know how to complete, and if you actually try performing the tasks on your test network. If this manual is too curt or hard to follow, you could try a more verbose book on vulnerability assessment, such as Hacking Exposed 4th Edition.
4. Now that you understand many of the general security ideas, it is time to get current. This is one area that has actually become easier in the last decade. The thinking used to be that vulnerability information should only be distributed to well-known and trusted administrators and security researchers through private digests such as Zardoz. This was a disaster for many reasons, and the full disclosure movement was born. In the last couple of years things have started to shift toward more limited ("responsible") disclosure and there is also a disturbing pay-money-for-early-disclosure trend. But information is still much more available than it used to be. Most of the news is carried on mailing lists, and I archive the ones I consider the best at Lists.Insecure.Org. You must subscribe to Bugtraq, and I would also highly recommend pen-test, vuln-dev, and security-basics. Read at least the last 6-12 months of archives. Choose other lists that correspond to your interests. SecurityFocus also offers a security-jobs list which is an excellent resource for finding jobs or just understanding what employers desire.

   There are two major reasons for reading Bugtraq. One is that you must react quickly to new vulnerabilities by patching your servers, notifying your clients, etc. You can get this by simply scanning the subject lines or advisory summaries for bugs that directly apply to you. But then you will miss out on another crucial purpose of Bugtraq. Actually understanding a vulnerability helps you defend against it, exploit it, and identify/prevent similar bugs in the future. When you are lucky, the advisory itself will provide full details on the bug. Check out this excellent recent advisory by Core Security Technologies. Note how they describe exactly how the Snort TCP Stream Reassembly vulnerability works in detail and even include a proof-of-concept demonstration. Unfortunately, not all advisories are so forthcoming. For bugs in Open Source software, you can understand the problem by reading the diff. The next step is to actually write and test an exploit. I would recommend writing at least one for each general class of bug (buffer overflow, format string, SQL injection, etc.) or whenever a bug is particularly interesting.

   Be sure to read the latest issues of Phrack and the research papers posted to the mailing lists. Send your comments and questions to the authors and you may start interesting discussions. Read well-regarded books on the security topics that interest you most.

   I can't emphasize enough that you should intersperse hands-on work with all of this reading. Install unpatched RedHat 8 (or whatever) and run Nmap and Nessus against it. Then compromise it remotely, maybe via the latest Samba hole. Start out with a prewritten exploit from Bugtraq, which isn't quite as easy as it sounds. You may have to modify the 'sploit to compile, brute force the proper offset, etc. Then break in again using a different technique, and your own exploit. Install Ethereal and/or tcpdump and ensure you understand the traffic on your network during both your exploitation and normal network activity. Install Snort on an Internet-facing machine and watch the attacks and probes you'll experience. Wander around your neighborhood with Kismet, Netstumbler, or Wellenreiter on your Laptop or PDA to look for open WAPs. Install DSniff and execute an active MITM attack on an SSH or SSL connection between two of your computers. Take a look at my Top 75 Tools List and ensure you understand what each does and when it would be useful. Try out as many as you can.

5. Take a vacation, or at least a weekend camping! You deserve it! The steps above would probably take at least 3-12 months full-time, depending on your motivation level and the depth and breadth of your research.

Step 2: Now apply your newfound knowledge

Now you have learned enough to be dangerous. At this point, you would have little trouble obtaining most certifications, after studying the specifics of each topic. If your main goal is to find a job quickly, perhaps adding these extra feathers to your cap might be worthwhile. But I think your best bet is to prove your knowledge by joining and contributing to the security community. While this does indeed help others, it isn't an entirely selfless act. It improves your skills, leads to important contacts, and demonstrates your knowledge and ability in a constructive way. The latter is important if securing a career is one of your goals. These steps should also be fun! If not, perhaps you should keep looking at other fields. Here are some ideas:

Start participating with insightful comment and answers on the mailing lists. This is very easy and serves as a great learning experience, way to meet people, and garners some name recognition. If a security manager with a stack of 60 resumes recognizes your name, that is a huge win!

When a new worm or a big new vulnerability comes out, everyone wants to know the details. If you stay up all night disassembling the worm/patch and write the first comprehensive analysis, many folks will find that valuable. And you will learn a lot. Let your first priority be quality - if someone beats you to it, just compare your results with theirs to see if you (or they) missed (or misinterpreted) anything. You can also post your own exploits, although that is more of a political hot potato.

Attending security conferences is a great way to learn, party with fellow hackers, and network (in every sense of the word). Much better is to speak at these conferences. This field changes rapidly so there are always new topics and technologies to discuss. You don't have to be a well-known expert with a long history - just learn your topic well and put in the effort for a quality presentation. You could present at Defcon, at one of the more commercial events, or at a smaller regional con like ToorCon, CodeCon, Hivercon, etc. Among other advantages (often free admission/travel/hotel), this is a great way to meet people with similar interests. I spoke at the latest CanSecWest and have submitted a proposal for the next Defcon.

Now that you've seen and understand a wide variety of software vulnerabilities from your Bugtraq research, start finding your own. You can start by downloading any PHP app from Sourceforge. Most of those are hopelessly vulnerable to Cross-Site-Scripting, SQL injection, and/or remote code execution by "remote include" directives. Many (if not most) Windows shareware daemons are also vulnerable to simple buffer overflows and format-string bugs. Notify the authors and then write an advisory. After a few of these "easy targets", try breaking some more widely deployed programs.

Write a security tool! I could list some suggestions, but by this point you will have many of your own ideas as to what is needed. Scratch an itch.

I hope this helps. If you want more suggestions, Ask Slashdot. From that story, I found this post particularly insightful, especially the emphasis on "people skills". I don't claim to have any, but understand the value :).

5) Have you ever been tempted to use your gifts...
by Tim_F

...in a negative manner?

Have you ever hacked into someone else's computer? Have you ever considered it? What would cause you to think of doing this? Would your tools (nmap, etc.) be enough to allow you to do this?

And if you haven't, why is that the case?

Fyodor

I never do script-kiddie style "hack any random vulnerable box on the Internet" cracking. But sometimes I will launch targeted attacks at specific companies. I'll usually start with just a web browser and various search engines to learn everything I can about my target. I need to understand what the company does, who it partners with, and whether it has any corporate siblings, subsidiaries, or parents. Beyond that, posts by individual employees can be a gold mine. Besides providing names and titles for social engineering and brute force password attacks, the IPs in the mail/news routing headers can be very valuable. One of the reasons I run my own mailing list archive is to maintain access to the raw mail folders which contain the routing info and X-no-archive posts that web archives strip out. Another advantage to locating employees is that you can send them trojan executable attachments, which can be a very effective way into the network.

Next I'll gather known IP network information on the companies via DNS, whois, regional registries like ARIN, routing info, Netcraft, etc. Then comes the scanning (I tend to use Nmap), application-probing, vulnerability discovery, and exploitation stages.

Of course, I only do this when the company is paying me to do so. Performing these pen-tests offers several advantages over blackhat activity:

1. You don't go to jail (If you've worded your contract carefully.)
2. Instead of having to keep your übertechniques secret to avoid prosecution, you get to demonstrate them to management.
3. They actually pay you for this! And you are helping to protect them and the privacy of their customers.

Now some people might ask how you gain these skills without practicing on other networks first. Cheap hardware and the evolution of free UNIX operating systems have made this much easier than in the past. See the previous answer for some suggestions. And remember that you can always work together with friends, or participate in hacking contests like Defcon's Capture the Flag.

6) You'll have seen a lot of breakins.
by Hulver

During your time running Honeypots, you'll have seen a lot of compromised systems. Is there any incident that's really stuck in your mind because of the audacity of the attempt, or the stupidity of the person attempting the breakin.

Fyodor

On the humorous front, one attacker was was running a public webcam during his exploits, so we were able to watch him crack into our boxes in real time :). I will resist the urge to link a screenshot. His rough location was determined when we noticed Mrs. Doubtfire playing on his TV and correlated that with public schedule listings. He was working with a Pakistani group, but was actually on the US East Coast.

In the "disturbing audacity" front, this year we found that a group of crackers had broken into an ecommerce site and actually programmed an automated billing-sytem-to-IRC gateway. They could obtain or validate credit card numbers by simply querying the channel bot! Expect a more detailed writeup soon.

7) What makes a honey net enticing?
by cornice

It seems that many of the honey nets that the average hobbyist would run are built to attract a lesser cracker. What I mean is that ports are left open that normally would not be left open. Services are running that normally should not, etc. I think that a really smart fish would see this as nothing but a cheap lure and refuse the bait. Do you think it's possible to fool the really smart fish? Is is possible to bait with something enticing enough without tipping off the big fish? Does publication of your work make this task more difficult?

Fyodor

Excellent question, and I had many of the same concerns upon joining the project. Then I remembered that most of the attacks and real-world compromises are committed by these marginally skilled script kiddies. So there is still a lot of value in understanding their tools, tactics, and motives. Despite this apparent limitation, I have been surprised by some of the sophisticated things we have found. For example, the first known "in the wild" attack using the Solaris dtspcd vulnerability was caught by one of our honeynets and resulted in this CERT advisory. Then one of our Honeynet Alliance members had their Win2K honeypot compromised and joined into a botnet with 18,000 machines! Attackers on such a grand scale won't even know all of the companies they have compromised, much less whether any of the systems are honeynets.

I do believe baiting the "smart fish" might be possible, but I have never done this. Is not legally entrapment, as we aren't any sort of police force, but I am not very comfortable with the idea. If someone attacks my box that is just unobtrusively sitting on the network, I believe the attacker should have no expectation of privacy for his activities on the system. Things become more complex if I try to lure the attacker.

8) IPv6
by caluml

Do you think that with the very large address space of IPv6 that random scanning for a certain port will die off? (I notice nmap doesn't support random IPv6 address scanning - maybe you've already come to the same conclusion?) Simply put, the chances of finding a machine if it's not advertised anywhere will be very much reduced. Will this make people lazy and complacent, trusting on the large numbers involved to protect them?

Fyodor

Finding a machine by by pinging a completely random 128-bit address will probably never be effective. Fortunately, we won't have to! Nmap does not even do that for 32-bit IPv4 addresses - it is smart enough to skip huge blocks of address space that are unallocated or used for private (RFC1918, localhost) addresses. We will also see patterns emerge for IPv6. For example, they may often be allocated sequentially so that finding one leads to many others. I am waiting until adoption rises and we start seeing these patterns emerge before I can implement them appropriately in Nmap. Certain new DNS features may also prove useful for locating IPv6 machines and networks.

9) standalones and small home nets
by zogger

it seems like most of the emphasis is on enterprise networks, but that still leaves millions and millions of home machines and small home networks just stuck. What do you see as some of the trends and solutions for those people? Their data and system integrity is just as important to them as any corporations is, and usually not having the appropriate skill set, is even harder to implement.

Fyodor

I am afraid the focus by security companies on enterprise networks will continue, as that is where the money is. The good news is that securing small home networks is far easier. But that doesn't make it simple, nor mean that many people will bother. I would categorize the risks into 3 categories:

Traditional network server vulnerabilities: Your average home user doesn't need to run any network daemons or have any TCP/UDP ports open to the Internet. Most of the time they only have 1 IP, used either by a standalone PC or a NAT device (e.g. "broadband router") in front of their small network. This is a good configuration, as it limits what attackers can reach directly. But you need to be sure that the IP doesn't have any unnecessary ports open. You can verify this by running 'netstat' on the Windows or UNIX machine using the IP. I would also recommend confirming using a port scanner such as Nmap. Here are example commands:

nmap -p- -sS -T4 -v -O [your IP]
nmap -p- -sU -v [ your IP ]

The TCP and UDP scans could be combined into one execution, but are listed separately since the TCP scan may go much faster. Remote UDP scans are also less reliable against some heavily filtered hosts. You may have to rely on the netstat info or configuration details in this case.

Any open ports found should be evaluated with extreme prejudice. Unless clearly necessary, close Windows file sharing, external NAT device admin ports, and everything else found.

Don't forget the wireless backdoor! Blocking the Internet link from your private machines is insufficient if anyone can hop on your open WLAN and attack your machines. WEP isn't perfect, but the 104-bit (so-called 128-bit) version should at least keep people from accidentally connecting to your network or sniffing your data. Be sure to set a good password and upgrade to recent firmware for your WAP and other network devices.

Subscribe to the security advisory lists for all the operating systems (and devices, if available) you run. Major vendors such as RedHat, Debian, FreeBSD, Mandrake, and Microsoft all offer these. Most even offer automatic updates if you desire that.

Client vulnerabilities: Once you close the services you don't need (ideally all of them), client vulnerabilities must be addressed. Keeping your web browser and mail reader up-to-date is particularly crucial. Also harden them as much as possible. For example, IE is full of holes but at least has a good interface for site-by-site security policies (Tools -> Internet Options -> Security). Go through and neuter the "Internet zone" settings by disabling ActiveX and Java. In the rare case that sites need this, find an alternative site or add them to the trusted zone. If your are really serious about security, neuter "trusted sites" and "local intranet" privileges as well. Many recent IE vulnerabilities trick the browser into using the wrong zones. Consider using a different browser. Also configure your mailer to disregard HTML and JavaScript.

Remember to pay careful attention to security warnings, whether they come from IE, Mozilla, your ssh client, or anything else. Don't just click OK. And don't shoot yourself in the foot when configuring your apps. It is hard to entirely blame the vendor when users tell P2P apps or Windows filesharing to share their whole drive without any password. Failing to change default passwords or enable basic restrictions on X Window or FTP servers is only slightly more forgivable. All of these errors happen frequently! The apps/devices should be secure by default, but you have the ultimate responsibility for protecting your data.

Malware: This is what I consider the biggest problem on desktops: people running applications they can't trust. Email borne viruses, worms and trojans are an obvious example. Be very careful what you click on. Unfortunately, it is very difficult to know what to trust. Mail is trivial to forge, and even the "proper" installers for many P2P applications infest your computer with loads of invasive spyware. Even Intuit TurboTax was caught writing to customers' boot information track.

What can you do? My honest suggestion is to run peer-reviewed open source applications on a free OS such as Linux or FreeBSD. You still have to be careful, but these problems are far less prevalent on UNIX platforms, which also have better tools and procedures to deal with them.

What if dumping Windows is not an option? Run NT/2K/XP instead of Win9X/ME, and try to run everything you can as an unprivileged (non-administrator) user. Be extraordinarily careful about what you install and run, and make frequent backups. You might also want to look into a personal firewall such as Zone Alarm (limited free version.

10) What is your favourite tool?
by Noryungi

I have just read your top 75 security tools list. Thank you for posting all this information, which I am going to study very carefully.

One question though: in all these tools, which one is your personal favourite? (This excludes Nmap, of course).

Fyodor

I have far too many favorites among this great group to choose just one! But here are a few developers and tools that are particularly worthy of mention:

One of the people I most admire in the security field is Solar Designer. He is a guru in networking, security, and low level kernel/assembly/architecture details. He has also created many tools that security professionals use daily. Yet he never exhibits the arrogance, elitism, and egotism that sadly characterizes so many "stars" of the security community.

Among SD's tools is John the Ripper, my longtime favorite local password hash cracker. It has been around forever, but was written with a flexible and powerful interface while keeping extensibility in mind. So it is still as useful in these days of shadowed password files and MD5/Blowfish hashes as it was back in the days of crypt() and unprotected /etc/passwd. Lately SD has been working on the Owl secure GNU/Linux distribution, which can be installed on disk for hardened systems like firewalls, or booted and run from CD as an easy way to run security tools such as John and Nmap.

Another of those "brilliant yet still nice" security developers is Dug Song. Even after the seminal "Insertion, Evasion, and Denial of Service" paper by Ptacek and Newsham, many IDS vendors continued to ignore the problem. When Doug released Fragrouter (now fragroute), which implements some of these attacks, vendors finally took notice! He has also written the excellent libdnet library, but my favorite of his tools is DSniff, a suite of tools for advanced network sniffing and "monkey-in-the-middle" attacks. It even handles ARP poisoning and other techniques for sniffing hosts on a switched LAN.

While I'm on this topic, let me also give "mad props" to the Hping2 packet prober, Kismet wireless stumbler, Ethereal packet decoder, Netcat, recent THC releases, Snort IDS, the Nessus vulnerability scanner, and all the other great Open Source tools out there!

I would also like to thank Slashdot for granting me this interview and to everyone who asked such excellent questions. I only wish I had time to answer more of them. Then again, I have probably rambled on enough. Now it is your turn to ramble in the comments :).

Cheers,
Fyodor

rajiv was among dozens to report that unlike most "Hacks" in film, The Matrix reloaded actually has an ounce of reality where other films would rely on fancy 3D graphics. You can see more at Insecure.org where they have screenshots. It's only on screen for a split second, but Tritnity uses Nmap to find a vulnerable SSH server, and then exploits it using the SSH1 CRC32 exploit from 2001.

Fyodor is the driving force behind Insecure.org and the top-rated Nmap network exploration and security auditing tool. He's also involved in The Honeynet Project (and is a coauthor of the project's book, Honeynet: Revealing the Security Tools, Tactics, and Motives of the Blackhat Community). One question per post, please. We'll run Fyodor's answers to 10 of the highest-moderated questions as soon as he gets them back to us.

Fyodor is the driving force behind Insecure.org and the top-rated Nmap network exploration and security auditing tool. He's also involved in The Honeynet Project (and is a coauthor of the project's book, Honeynet: Revealing the Security Tools, Tactics, and Motives of the Blackhat Community). One question per post, please. We'll run Fyodor's answers to 10 of the highest-moderated questions as soon as he gets them back to us.

spring writes "Every so often, the author of everyone's favorite network reconnaissance tool, nmap, runs a survey to determine which security-oriented software products are most popular. This year's tool survey was just released, and it contains some interesting results. Old favorites like Nessus, Snort, Netcat, and Ethereal made the list, of course. SAINT and SARA are still around. But a number of new tools appeared this year, like Windows-only GFI LANguard, SuperScan, and Cain & Abel. Nikto and Kismet demonstrate the growing importance of wireless networks. The survey contains many good tools. Certainly worth a read."

spring writes "Every so often, the author of everyone's favorite network reconnaissance tool, nmap, runs a survey to determine which security-oriented software products are most popular. This year's tool survey was just released, and it contains some interesting results. Old favorites like Nessus, Snort, Netcat, and Ethereal made the list, of course. SAINT and SARA are still around. But a number of new tools appeared this year, like Windows-only GFI LANguard, SuperScan, and Cain & Abel. Nikto and Kismet demonstrate the growing importance of wireless networks. The survey contains many good tools. Certainly worth a read."

spring writes "Every so often, the author of everyone's favorite network reconnaissance tool, nmap, runs a survey to determine which security-oriented software products are most popular. This year's tool survey was just released, and it contains some interesting results. Old favorites like Nessus, Snort, Netcat, and Ethereal made the list, of course. SAINT and SARA are still around. But a number of new tools appeared this year, like Windows-only GFI LANguard, SuperScan, and Cain & Abel. Nikto and Kismet demonstrate the growing importance of wireless networks. The survey contains many good tools. Certainly worth a read."

scubacuda writes "According to Wired, the University of Toronto's Internet Censorship Explorer permits people test the limits of national and organizational Internet-blocking schemes. Users enter a target URL (and a country), and the software then scans the ports of available servers in that country, looking for open ones to connect on from behind that country's firewall. Many consider port scanning a gray area, as it's often used by various hackers to find vulnerabilies that can be exploited."

taviso writes "Are you bored with your OS fingerprint? Do you dream of being able to impress your friends by convincing them your webserver is running on a sega dreamcast, or Apple LaserWriter? Well Dream no more! David Berrueta has written a paper oulining the techniques and tools available to defeat nmap's OS fingerprinting, available here [pdf]. Besides the hours of entertainment this could provide, he also lists some of the more serious reasons why you might want to consider this."

taviso writes "Are you bored with your OS fingerprint? Do you dream of being able to impress your friends by convincing them your webserver is running on a sega dreamcast, or Apple LaserWriter? Well Dream no more! David Berrueta has written a paper oulining the techniques and tools available to defeat nmap's OS fingerprinting, available here [pdf]. Besides the hours of entertainment this could provide, he also lists some of the more serious reasons why you might want to consider this."

Effugas writes "After pushing OpenSSH to perform feats of secure tunneling far beyond what I ever expected it could do, it became clear that some genuinely useful modes of network operation were simply inaccessable without either replacing or manipulating core network protocols. Since the basic infrastructure of the Internet isn't likely to change any time soon, that left...creative manipulation and reconstruction of the Lingua Reseaux: TCP/IP. Taking advantage of expectations, pitting layers against eachother, finding new uses for old options and data fields -- instead of simply unleashing the latest incarnation of some "Ping of Death", could such work unveil hidden functionality within existing networks? As I discussed at Black Hat 2002 and the inimitable Defcon X, the answer is yes. And now, proof of this is ready. BSD Licensed (in deference to the very source of TCP/IP), The Paketto Keiretsu, Version 1.0, is a collection of five interwoven "proof of concepts" that explore, extract, and expose previously untapped capacities embedded deep within networks and their stacks, at Layers 2 through 4. The five -- scanrand, minewt, lc ( linkcat ), paratrace, and the OpenQVIS cross-disciplinary-a-go-go phentropy -- demonstrate Stateless TCP Scanning, Inverse SYN Cookies, Guerrila Multicast, Parasitic Tracerouting, Ethernet Trailer Cryptography, and quite a bit more. (For details, stop by DoxPara Research or check out the latest slides. The academic paper is coming "soon".) In terms of actual usefulness, scanrand is no nmap, but it's still interesting: During an authorized test inside a multinational corporation's class B, scanrand detected 8300 web servers across 65,536 addresses. Time elapsed: approximately 4 seconds."

Istealmymusic asks: "Now that Nmap 3.00's idlescan technique is fully documented, thousands of vulnerable NT and Linux hosts on the Internet are being exploited to perform stealthy port-scanning. My employer's Linux cluster was a victim of these attacks; apparently he has been used to perform hundreds of port scans on DDN machines. Needless to say we where contacted by the sysadmin and forced to blacklist the cracker. However, our Linux cluster is still vulnerable to the idlescan exploit from other attackers, and I believe our company has a false sense of security. OpenBSD is the only OS I know of which randomizes the IPID sequence therefore making it invulnerable to the idlescan, but we have neither the time nor urge to migrate to OpenBSD. How can one secure their Linux or NT TCP/IP stack from malicious idlescanning?"

Istealmymusic asks: "Now that Nmap 3.00's idlescan technique is fully documented, thousands of vulnerable NT and Linux hosts on the Internet are being exploited to perform stealthy port-scanning. My employer's Linux cluster was a victim of these attacks; apparently he has been used to perform hundreds of port scans on DDN machines. Needless to say we where contacted by the sysadmin and forced to blacklist the cracker. However, our Linux cluster is still vulnerable to the idlescan exploit from other attackers, and I believe our company has a false sense of security. OpenBSD is the only OS I know of which randomizes the IPID sequence therefore making it invulnerable to the idlescan, but we have neither the time nor urge to migrate to OpenBSD. How can one secure their Linux or NT TCP/IP stack from malicious idlescanning?"

Anonymous Coward writes: "The Journal of Technology Law and Policy has a good article on computer security and privacy. If you ignore the more metaphorical crap at the beginning of the article, the author marches through some laws that apply to the Internet and shows how they apply and why his way of deciding what kind of access to a computer breaks the law and what kinds don't is better. (Its based on property and expectations of privacy.) It's interesting to see the computer security from a lawyer's point of view. Especially interesting are his claims that using nmap is illegal, despite the VC3 v. Moulton case. I'm not sure I agree with him, but he definitely makes a pretty sobering case." Actually, I think the metaphors throughout this piece (not just at the beginning) are what make it interesting, and a big component of law is dealing with metaphors. This piece also collects in one place a lot of the cases dealing with computer law.