Slashdot Mirror


Ethereal Packet Sniffing

nazarijo writes "I look at packets for a living. I generate them, I capture them and dissect them, and I try and make sense of them as quickly as possible. Sniffers and protocol analyzers are part of my bread and butter, and I'd be foolish to not use Ethereal. Tcpdump for a quick capture, but I use Ethereal when I need detailed information in a better, more navigable fashion. Because of that, I was pretty interested to see a book on Ethereal coming out." Read on for Jose's review of Ethereal Packet Sniffing from Syngress. Ethereal Packet Sniffing author Angela Orebaugh with Greg Morris and Ed Warnick pages 468 publisher Syngress rating 7 reviewer Jose Nazario ISBN 1932266828 summary Solid coverage of an excellent networking tool. Offers value beyond free documentation, insight available nowhere else, and plenty of handy tips and tricks.

I've used the tool for years, and I've read the docs a bit, so I felt comfortable with the tool. Still, I wanted to learn something new with it, and I wanted to see if this book could offer what I was hoping for. The book delivers, and does a pretty good job. One of the big tests for me about any book that covers an Open Source project is "Does this book offer more than the existing documentation?" If it fails to, the book isn't worth the money, I'll stick with free docs. While the book comes out favorably for me, I'll start with the things I didn't like, first.

One of the big things that is missing from this book is any coverage of Ethereal on OS X. Given how many people are migrating to OS X (from UN*X or from Windows), and the coverage of Ethereal on Windows, I would have expected some mention of it. Luckily it's available in both Darwin Ports and the Fink project, but some mention of any of the quirks people may encounter would have been welcome. Amy (from Syngress) tells me that they will have a paper in their Solutions center on Ethereal on OS X, which would be great to see.

Another annoyance with the book is the repeated coverage in some sections of various aspects of Ethereal. One that stands out is the coverage of the additional tools which are installed alongside Ethereal, like Editcap and Text2pcap. They are covered in chapter 2 for a bit and then more completely in chapter 6. Covering these tools only once would have sufficed, but it does let chapter 2 stand on its own. Amy tells me that they do this intentionally, because it makes some chapters stand on their own as "units" for others to use. That makes sense.

A final bit of the book I didn't like was the choice of screenshots: quite a number of the screenshots were full screen dumps when only one or two elements of the page really mattered. Either trimmed or annotated screenshots would have been more welcome. A lot of information gets dumped in Ethereal, helping people navigate the UI with a static, black-and-white image would have been welcome.

Now, on to the real strengths of the book. Like I said earlier, The book offers more coverage than the existing, free docs on Ethereal provide, or at least in a more manageable form. Obviously, with the source code in front of me I could dissect the tool and learn everything about it, but that's hardly efficient. Simply put, the book introduces network sniffing and troubleshooting well. How can you place a sniffer to get coverage, what can a sniffer tell you during troubleshooting (and what can it not?), and of course how to get and install Ethereal (on UN*X and Windows).

The next chapter covers exactly what you would expect it to, how to use Ethereal. Ethereal's main use is as a GUI protocol analyzer, so you have menus, panes and windows to navigate. This chapter tells you what they are and how they present and format the data you're looking at. The next chapter deals with four tools that come with Ethereal: Tethereal (very similar to tcpdump), Editcap, Mergecap, and Text2pcap (all useful for managing pcap files).

Chapter 7 is one of those handy things to read. Ethereal is typically used to read pcap files, but it can also read snoop files, Microsoft Network Monitor files, EtherPeek files, NAI's Sniffer files, and HPUX's nettl files, all of which you'll find around. It's handy that you can see how to integrate Ethereal with these other products.

Chapter 8 brings it all together with real world packet captures, many of which are also on the included CD. These files include scans, Trojan uses, and even worm traffic. All of these are useful for learning how to use Ethereal and highlight the power of the tool. You can go from novice to a pretty decent network protocol junkie if you dilligently study the resources in this chapter and on the CD.

Chapter 9 will be useful to a small subset of people, but quite useful. This chapter gives you a tour of how to develop for and extend Ethereal. Ethereal's main strength is a huge number of decode routines, such as sFlow and MPLS (in addition to the standard ones like DNS, DHCP, and the like). Using this information you can extend Ethereal for your own needs and maybe even contribute back to the project.

Either the developer's angle or the detailed discussions and examples of the filter syntax are my favorite parts of the book. They contribute significant value for everyday use, and I found them useful in a recent task at work.

The book is going to run the risk of becoming quickly out of date, given the development pace of Ethereal. However, it relies more on underlying core concepts and principles inherent in Ethereal, so it should stay useful for longer than you may think.

All in all I would say this is probably worth picking up if you're looking at becoming a network operator or network security junkie. You'll learn a lot about a powerful tool, how to integrate it into your use, and even how to dissect real traces of traffic. I give it a 7 out of 10 for the above weaknesses, but that shouldn't stop you from strongly considering it.

You can purchase Ethereal Packet Sniffing from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page

22 of 147 comments (clear)

  1. Additional note: by Anonymous Coward · · Score: 5, Funny

    I purchased this book using credit card information I picked up using Ethereal.

    1. Re:Additional note: by Anonymous Coward · · Score: 3, Funny

      I should also thank you for not posting a joke like:
      In Soviet Russia Ethereal you!
      or ...
      Ethereal
      ???
      Profit!

      There! I just earned my second "Troll" mod for the week.

      --
      Getting modded as "Troll" let's me know that I'm not succumbing /. groupthink.


      Getting modded as "Troll" lets everyone else know your an idiot.

    2. Re:Additional note: by Otter · · Score: 3, Funny
      Getting modded as "Troll" lets everyone else know your an idiot.

      If you're going to be an apostrophe troll, you need to make sure you get your own apostrophes right!

  2. I love this tool by jxs2151 · · Score: 3, Funny

    Can only understand about half of what it does though. Maybe I'll buy the book.

  3. Already out of date by Anonymous Coward · · Score: 5, Informative

    While this is an interesting book, its problem is that it is already out of date. It seems that it was written at a time when the user interface was undergoing lots of churn.

    For example, on page 47, figure 2.1 is out of date, as the menu items have changed and the toolbar now has more items.

    On Page 146 and 147 the authors attempted to deal with changes in the GUI, and show us what the new print dialog box will look like, however, that version is also out of date.

    On Page 153, Figure 4.19 is out of date. On Page 155, Figure 4.21 no longer exists. Page 156, Figure 4.22 is out of date. Page 162, Figure 4.31 is out of date, and so on.

    Further on, Page 180, Figures 4.49 and 4.50 are also out of date, and it would have been nicer to show some real-life examples of problems one can spot with the Time Sequence Graphs and some explanation of how socket layer stuff relates to what you might see on the wire.

    So, I am not sure this book is worth buying. Perhaps wait for the update.

    1. Re:Already out of date by Wister285 · · Score: 3, Informative

      I'm pretty sure this is why they offer a "1 Year Upgrade", as shown on the cover. It is meant to protect the customer against these sorts of problems with rapidly progressing technology. The only trick seems to be that you have to register. Still, it's refreshing to see book publishers offering such things.

    2. Re:Already out of date by ted_nugent · · Score: 3, Interesting
      Sounds great until you realize that Syngress has never published a revision to any of their existing titles.

      Disclaimer:Of course, I could be totally wrong on this, since it's based on my own casual observation rather than a publishing schedule.

      --

      Free the West Memphis Three!

  4. Just Wondering.... by Dr.+Bent · · Score: 4, Funny

    I look at packets for a living. I generate them, I capture them and dissect them, and I try and make sense of them as quickly as possible.

    So what's it like working for the N.S.A.? Do they have a decent benefits package?

  5. Question... by Frennzy · · Score: 4, Interesting

    Can we assume that it really focuses more on the ethereal product than analyzing and understanding frames? (In short, is it more for someone who wants to squeeze the most out of ethereal, or does it do remedial to advanced instruction on packet construction, deconstruction, and analysis?

  6. I'd love to but... by Iscariot_ · · Score: 4, Interesting

    I'd really love to play around with Ethereal, but I'm running WindowsXP and for some reason it just doesn't go. I've read that this has to do with WinPcap.

    What I want to know is, is there a way to get Ethereal running on XP? Is there an alternative to WinPcap 3.0?

    1. Re:I'd love to but... by lukewarmfusion · · Score: 4, Informative

      I've run Ethereal on several XP boxes. Make sure you install WinPCap first. Check which device you have set to monitor, too.

      Your network configuration can also affect what packets you see - are there switches dividing your network? Are you alone on your network?

      New the Ethereal?
      Start a capture, then check your email. Then use the email address and password you capture to do all kinds of nasty things.

    2. Re:I'd love to but... by fred_sanford · · Score: 5, Informative

      Give credit here but this is how I got it working on 2000:

      [quote] For windows get winpcap [polito.it]
      then get ethereal for windows [ethereal.com]
      and get windump [polito.it]

      SANS.org has all the info: Packet capture apps [sans.org][/quote]

    3. Re:I'd love to but... by phaetonic · · Score: 4, Insightful

      Why not get Knoppix or one of the many LiveCD distributions that allow you to use ethereal without the need to install Linux. Then your WindowsXP problem will not longer be a problem. The Knoppix release I used last July had support for my wireless NIC (Orinoco) on my laptop, and everything worked great.

  7. OS X & Ethereal by grocer · · Score: 3, Insightful

    Ethereal requires X Windows to run on OS X...which means the some form of a rootless install or the defunct Apple XFree86 Beta in Jaguar (10.2.x).

    Panther (10.3.x) has X Windows intergrated, although I haven't bought it yet...(so I don't know how well it works or if all the build issues are sorted out of Fink...although Fink is supposed to work now)

    10.1.x, I have no clue, but it's different than 10.2.x (probably have to install some third variant of X via Fink)

    Ethereal on OS X does rock especially with KisMac but there's three or four possible scenarios for install...probably why the book doesn't cover it...

  8. Great every idiot on slashdot will be sniffing now by DR+SoB · · Score: 3, Funny

    Here we go with the n00b questions.. ie. Can it sniff spam packets? Answer: No, spam packets are so mysterious and powerful, no available NIC is capable of passing them to a sniffer program.

    Please people, leaving SNIFFING to the professionals!

    --
    Mod +5 Drunk
  9. Re:possible? by SquadBoy · · Score: 3, Funny

    No spam packets are unlike other packets. They are marked with the "spam bit" and this means that sniffers will not capture them or display them.

    --

    Cypherpunks: Civil Liberty Through Complex Mathematics. Those who live by the sword die by the arrow.
  10. Ethereal in University Setting by crass751 · · Score: 4, Interesting

    In the networking class I'm taking this semester, we've been doing exercises using Ethereal to study different protocols and layers of the TCP/IP stack. My professor is working on a book that uses Ethereal to study networks, but provides all the relevant captures and such to keep students from running traces on active networks. It's been a useful learning aid, for me at least. It's makes more sense to think about packets and such when you can actually see them and the data they contain.

    1. Re:Ethereal in University Setting by dr_dank · · Score: 3, Funny

      My professor is working on a book that uses Ethereal to study networks, but provides all the relevant captures and such to keep students from running traces on active networks

      Is this Prof on crack that he/she doesn't think that any of their students is going to try sniffing their neighbors packets on the dorm network? Hell, thats the first thing I'd do!

      --
      Where does the school board find them and why do they keep sending them to ME?
  11. packet sniffing by Cruciform · · Score: 4, Funny

    Ever mention 'packet sniffing' in a public place?

    Suddenly people across the room are hanging on your every word, until they realize you didn't say "panty sniffing" and they can't get vicarious thrills/outrage from the perverted geeks in the corner.

  12. Flameon by g0bshiTe · · Score: 3, Funny

    Not to mention the 13 root exploits for Ethereal.

    --
    I am Bennett Haselton! I am Bennett Haselton!
  13. Usually the wrong level for solving that problem by billstewart · · Score: 3, Informative
    Spam doesn't arrive in packets - it arrives in SMTP sessions, packaged in TCP flows, packaged in IP packets. That means that you don't have a whole spam session in any given IP packet, so it's much harder to detect spamminess from sampling IP packets than from sampling at the SMTP handler. For most people, all the incoming SMTP is handled at one place - one of your home machines if you're at home, or one of your servers (or clusters of servers) if you're a mailbox provider or a business.

    If you're an ISP or hosting center that has customers that you're only providing with IP services, not email services, you _could_ sniff packets and send RSETs to kill sessions that look like spam, but you'd be doing it with less information than your customers, and you would probably end up killing off lots of useful mail, such as the message they're sending to abuse@example.net telling them how to find the spammer that just sent them this message. Usually a bad idea.

    --

    Bill Stewart
    New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
  14. Mordekainen's Ethereal Packet Sniffing by quantaq · · Score: 3, Funny

    Yes, but what are the material componants? Area of effect? Your review of this "Ethereal Packet Sniffing" leaves me wanting.

    Why oh why did I have to play so much D&D in high school...