Slashdot Mirror
Foundations of Mac OS X Leopard Security
jsuda writes "At least a half-dozen times in the book 'Foundations of Mac OS X Leopard Security' the authors state that there is a misconception that the Macintosh computer is immune from security problems. That allegation may explain why there are very few books published (and nearly none in recent years) about security for the Mac. This book is meant to change all that. The authors acknowledge that the Mac OS X software has had little of the security problem experience of Windows (and other operating systems, to a lesser extent) but they spend 455 pages detailing exactly where and how the Macintosh platform is (or may be) vulnerable." Read below for the rest of Jsuda's review.
Foundations of Mac OS X Leopard Security
author
Charles S. Edge, Jr., William Barker, and Zack Smith
pages
455
publisher
Apress
rating
9
reviewer
jsuda
ISBN
978-1-59059-989-1
summary
Best book on Mac Security
Many of the security issues raised in the book are theoretical or deal with added elements of the Mac software install that contain non-Apple components — Apache Web server and Perl and PHP scripting packages, for example. Many of the items of concern deal with generic problem areas of computer usage in general, both software and hardware, which affect the Mac as well as any other computers and networks. While the perspective of the book is on the Mac, much of the security review will apply to any type of computer or network.
Messieurs Edge, Barker, and Smith are seasoned Mac and security professionals who point out in a very systematic and comprehensive way the potential problems of running the Mac both in single use and networked environments. The focus is primarily on Mac OS X Leopard and the other software which comes with any new Mac computer, although there is some discussion of earlier OS X versions and earlier generations of Apple applications like Airport.
The book has five main parts covering general security matters, essential security fundamentals, networking, sharing, and workplace security issues. There are four very short appendices of modest value.
The initial first three chapters deal with general security and security fundamentals is basic stuff discussing how technical computer security issues are entwined with practical realities of using computers in a business or home, and that compromises between security and practicality generally must be made. There is discussion of types of security attacks, how the Windows booting programs, Parallels and Boot Camp, implicate Windows security issues on the Mac, and how the UNIX underpinnings of the Mac OS X allow for more sophisticated techniques and tools in securing the Mac computer and networks. Chapter 1 is a useful "quick start" guide of items which can be addressed readily by nearly any level of user to safeguard the Mac from many security concerns. Apple has provided a lot of built-in security features and services which can be adjusted by individual users to his or her own needs, like FileVault, Secure Trash, Keychain, permissions, and others. Higher-level users and maybe experienced security professionals not used to the Mac may be bored with the first part of the book.
Part two deals with protecting the Mac from malware and exploitable services in the OS and major applications like the Safari browser and Mail applications. It explains how malware can affect the Mac through script viruses, social engineering techniques, and other exploits. The book lists a number of available software tools which can help solve some of the potential problems. The section on reviewing and configuring monitoring processes and logs is especially interesting.
Securing networks, using and configuring firewalls, and wireless networking make up the bulk of part three. The content in chapters 7 through 9 is quite technical covering types of networks; routers, hubs and switches;proxy, DMZ, and other servers and hardware setups, advanced firewall configuration using both GUI and command line interfaces; filtering; traffic throttling; and more. The sections describing testing of firewalls and hacking wireless networks using tools like Kismac and iStumbler are especially useful.
Chapter 11, in part four, dealing with website security when utilizing the built-in Apple web services, includes a checklist of at least a dozen items to be dealt with in locking down a site. Security for remote conductivity is addressed also, with particular emphasis given to VPN, secure shell, and the use of network administration tools like Timbuktu and DAVE. Attention is given to both the standard Mac OS X installation as well as to OS X Server. The most complex discussions involve using Open Directory in a security plan. My favorite sections were in chapters 14 on network scanning, monitoring, and intrusion prevention tools. The book describes how to understand your own machine/network security status by learning how to attack other networks. And how to use techniques like white/black box testing, fingerprinting, enumeration, port and TCP/UDP scans, ping sweeps, and more.
The book describes how intrusion detection is accomplished. Guidance is provided on software tools like Tripwire, snort, Checkmate, and others. The last chapter concerns forensics and how to handle attempted or successful intrusions to both understand security weaknesses and to preserve evidence for civil or criminal proceedings, CSI-like.
Nearly all of the presentations cover two levels of interactivity using either GUI-based tools or the command line. Except for a handful of sections, the presentations are useful even for higher-end users, including those dealing with medium to large networks.
The writing is workmanlike and without style or wit, but carefully organized and expressed. There are plenty of (grayscale) screenshots of relevant software application configurations, and sidebar Notes and Tips on many topics. Anyone who has a serious interest in Mac OS X security will benefit from this book as its main virtue is its systematic and comprehensive approach to the issues. It is designed to inform users of all levels how and why to think about OS X security. Geeks who want or need to know Mac OS X security will get a nicely organized book sufficiently filled with useful content. This is not a book intended to raise all security issues or to provide all the answers. It does answer many problems, and will point nearly all users in the right direction for their specific needs.
You can purchase Foundations of Mac OS X Leopard Security from amazon.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.
Messieurs Edge, Barker, and Smith are seasoned Mac and security professionals who point out in a very systematic and comprehensive way the potential problems of running the Mac both in single use and networked environments. The focus is primarily on Mac OS X Leopard and the other software which comes with any new Mac computer, although there is some discussion of earlier OS X versions and earlier generations of Apple applications like Airport.
The book has five main parts covering general security matters, essential security fundamentals, networking, sharing, and workplace security issues. There are four very short appendices of modest value.
The initial first three chapters deal with general security and security fundamentals is basic stuff discussing how technical computer security issues are entwined with practical realities of using computers in a business or home, and that compromises between security and practicality generally must be made. There is discussion of types of security attacks, how the Windows booting programs, Parallels and Boot Camp, implicate Windows security issues on the Mac, and how the UNIX underpinnings of the Mac OS X allow for more sophisticated techniques and tools in securing the Mac computer and networks. Chapter 1 is a useful "quick start" guide of items which can be addressed readily by nearly any level of user to safeguard the Mac from many security concerns. Apple has provided a lot of built-in security features and services which can be adjusted by individual users to his or her own needs, like FileVault, Secure Trash, Keychain, permissions, and others. Higher-level users and maybe experienced security professionals not used to the Mac may be bored with the first part of the book.
Part two deals with protecting the Mac from malware and exploitable services in the OS and major applications like the Safari browser and Mail applications. It explains how malware can affect the Mac through script viruses, social engineering techniques, and other exploits. The book lists a number of available software tools which can help solve some of the potential problems. The section on reviewing and configuring monitoring processes and logs is especially interesting.
Securing networks, using and configuring firewalls, and wireless networking make up the bulk of part three. The content in chapters 7 through 9 is quite technical covering types of networks; routers, hubs and switches;proxy, DMZ, and other servers and hardware setups, advanced firewall configuration using both GUI and command line interfaces; filtering; traffic throttling; and more. The sections describing testing of firewalls and hacking wireless networks using tools like Kismac and iStumbler are especially useful.
Chapter 11, in part four, dealing with website security when utilizing the built-in Apple web services, includes a checklist of at least a dozen items to be dealt with in locking down a site. Security for remote conductivity is addressed also, with particular emphasis given to VPN, secure shell, and the use of network administration tools like Timbuktu and DAVE. Attention is given to both the standard Mac OS X installation as well as to OS X Server. The most complex discussions involve using Open Directory in a security plan. My favorite sections were in chapters 14 on network scanning, monitoring, and intrusion prevention tools. The book describes how to understand your own machine/network security status by learning how to attack other networks. And how to use techniques like white/black box testing, fingerprinting, enumeration, port and TCP/UDP scans, ping sweeps, and more.
The book describes how intrusion detection is accomplished. Guidance is provided on software tools like Tripwire, snort, Checkmate, and others. The last chapter concerns forensics and how to handle attempted or successful intrusions to both understand security weaknesses and to preserve evidence for civil or criminal proceedings, CSI-like.
Nearly all of the presentations cover two levels of interactivity using either GUI-based tools or the command line. Except for a handful of sections, the presentations are useful even for higher-end users, including those dealing with medium to large networks.
The writing is workmanlike and without style or wit, but carefully organized and expressed. There are plenty of (grayscale) screenshots of relevant software application configurations, and sidebar Notes and Tips on many topics. Anyone who has a serious interest in Mac OS X security will benefit from this book as its main virtue is its systematic and comprehensive approach to the issues. It is designed to inform users of all levels how and why to think about OS X security. Geeks who want or need to know Mac OS X security will get a nicely organized book sufficiently filled with useful content. This is not a book intended to raise all security issues or to provide all the answers. It does answer many problems, and will point nearly all users in the right direction for their specific needs.
You can purchase Foundations of Mac OS X Leopard Security from amazon.com. Slashdot welcomes readers' book reviews -- to see your own review here, read the book review guidelines, then visit the submission page.
OSX is intrinsically far more secure than Windows, but all systems have their vulnerabilities. For Mac pros to acknowledge that "far more secure" does not equal "completely secure" is a good step in the right direction. Thanks for the review, jsuda.
If you haven't been down-modded lately, you aren't trying.
Sacred cows make the best hamburger.
Sure they have, its just not a very useful platform to write viruses for since they have such a tiny market share. Hackers for hire use the ideal of "most bang for the buck" style so....windows it is. Turn the tides on market share and I bet you'd see a ton more viruses for OSX than you do now and it probably would be the Windows users saying....looks at all those viruses for OSX...their security sucks.
Insert funny smart-ass comment here.
Good question. But since servers tend to be protected a bit more than your average home users computer its a bit easier to get 100K of those than 1000 servers. But on very rare occasions a hacker figures out how to have his cake and eat it too.....
Insert funny smart-ass comment here.
That allegation may explain why there are very few books published (and nearly none in recent years) about security for the Mac.
I would think the reason is more that almost any book on UNIX security gets you 99% of what you need to know, and there are online sources to cover the rest.
Not that a book is not a good thing to see, but to my mind among admins or more serious users of OS X, the misconception that OS X is totally secure is in itself a misconception. OS X know systems will have vulnerabilities, but we also know there have been basically no attacks in the wild and that by default many things which might leave un-noticed holes (like web servers) are off by default - and that helps a lot, for the eventuality of real attacks coming someday.
To my mind, another aspect stopping attacks is actually the switch to Intel. That reset the counter for when we might see OS X attacks since buffer overflow stuff can't rely on which architecture it might hit. That and a more friendly update model (than Windows) that people actually apply when updates come.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
I don't think that's it at all. It's there is very little market for OS X security books at this point. Most people don't care. Let me explain.
On the home end of things, Macs are great and relatively secure. They do fine. That said, how many people buy books on Windows Security for those home computers? I'm going to say very few. Most people don't care or don't know they should do something to increase security.
The other front is businesses. Most businesses don't use Macs, by a large margin. Macs have a smaller enterprise market share than overall market share. If you are asked to secure a server or desktop, chances are it will be Windows or Linux.
These kind of books are, for the most part, targeted at administrators, businesses, etc. Since that market (administrators of Macs) is so small (compared to administrators of Windows boxes) there are very few books written.
This is compounded by the most important boxes to secure: web facing boxes (like servers). OS X Server's market share is very tiny compared Windows and Linux.
The books aren't there because the demand for them isn't very big, not because Mac users are think they are invulnerable from arrogance.
Comment forecast: Bits of genius surrounded by a sea of mediocrity.
Sure they have, its just not a very useful platform to write viruses for since they have such a tiny market share
There are now tens of millions of macs being used now. That's active use, not just purchased...
Now you tell me how in this day and age where viruses are all about building up botnets which are then sold, that a fairly homogenous systems with MILLIONS of systems to be had, is not a juicy target?
Marketshare alone is meaningless as a reason not to write viruses when you get to those kinds of numbers.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Take 'em all down, Fanboi! Good dog!
And with that message, your contentless response to a well-written message puts on display the fullest measure of your intelligence.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Gah.
A zombie can scan IPs for known security holes, but the programmer still has to design a virus, worm, etc. specifically for the Mac that will exploit that hole. The viruses that attack windows won't work on Mac...you have experiment and find different avenues of attack...identify the security hole to exploit. I maintain that few security holes have been identified because fewer people are looking for them, not because there are fewer of them.
Remind me again what makes the Mac very secure in an absolute sense? How do you measure it. Because the some of the well known vulnerabilities in other systems don't exist in the Mac?Are you saying the Mac doesn't receive email attachments, has a 100% secure browser, and isn't suseptible to worms?Removing avenues increases security, but it doesn't mean that a completely different system doesn't have different avenues of attack that don't exist in the competition. If you have an existing product, and close up one avenue of attack, odds are good that it hasn't opened up a new one, and the newer version will be more secure. When you build a different product, you can look at the mistakes made elsewhere and correct for them, but it's a pretty good bet that you've had the oversight or mistake elsewhere in your own design.
And it still seems like you're implying that the people at Apple have somehow come up with a brilliant way of stopping all attacks by worms, email attachements and browser exploits.
All I'm saying is that Mac's current security is due to market share, not intrinsic to design. Look, I'm a Mac user, too, but don't blind yourself or fool yourself into thinking that somehow those folks at Apple are geniuses who've made the perfect impenetrable system, while the dunces in Redmond couldn't figure out how to lock a door.
Well...
Mac OS X has some advantages in security. But I can't really say those advantages are due to Apple being somehow inherently 'better' coders than Microsoft or having made some kind of perfect system.
In my opinion, Mac OS X is less vulnerable than Windows in many areas is due to Apple being willing to go 'okay, this particular technology is dead, move along.' Microsoft relies on backwards compatibility for large market share; break backwards compatibility, and people do not upgrade. (Case in point: Vista.)
Apple has a smaller market share (and speaking as a Mac user and developer, we tend to sort of go, 'yes please, whatever you say' when they want to change things). We might bitch about it periodically (whither thou, 64-bit Carbon?), but this gives them the freedom to throw out legacy code and simplifies the code maintenance.
Or, in short: Apple's coders are not inherently better, but they end up with less old cruft to support and try to be aware of.
As a case in point, I'll note that the worst offender in terms of security on Mac OS X has, historically, been Quicktime. Quicktime is perhaps the oldest, most legacy-laden bit of crud in Apple's library. (The Quicktime APIs are darn near prehistoric, especially compared to things like CoreImage et al.) One would assume this means that Quicktime, more than almost anything else, has chunks of code that predate most of the programmers working on it, and which no one remembers or thinks about.
And in my experience, that's often where those kind of bugs come from... you change something, add a new bit of code that passing something into a function somewhere, completely unaware that four levels deeper there's some function which assumes the buffer is only 4k long. The old programmers knew there was an implicit limit down in this ancient routine, but no one now knows of that limit, and so -- unaware of this lurking nightmare 5 levels deeper in the stack -- they pass in a 6k buffer. Boom, security issue.
Windows has this problem in almost every corner of the OS. Worse, they cannot readily get around it... you can't just rewrite things from scratch, or you break legacy support! But as a result, there often are quite a few lurking behaviors that newer coders aren't aware of somewhere deeper in the system, things that never got documented, and which will eventually reach out to bite them.
Sure, there's situations which are just plain dumb (the carpet-bombing attack, for instance, is inexcusable behavior on IE's part), but most of those seem to be the minority.
So, yeah, Mac OS X has some advantage, as they have less legacy stuff to deal with. But even with that sort of advantage, no operating system -- not even Linux! -- is completely free of all flaws. We as users need to accept as a given that almost nothing is completely secure (at least, not and still be usable). This is especially true when many viruses and trojans rely on social engineering.
Even if Mac OS X prompts the user before allowing a program to elevate privileges, does that matter if users just click without looking? After all, lots of programs prompt for such things in order to install some shared framework they use at the installation or first-run stage. I know a lot of Mac users who just click on that warning blindly. And the warning doesn't matter if the user doesn't really pay attention.
So, yeah. Mac OS X may have less tangled, jungle-like legacy code for scary security holes to lurk in, but that does not mean it is invulnerable. Certainly not immune even to automated bugs, and especially not immune from social engineering.
Because the biggest security hole -- on ANY operating system -- is often user behavior.
There's my $0.02, anyway. :)
--Rachel
Back when Windows was first designed Microsoft assumed there was only one user on the computers and there was not way to get at the computer other than via the keyboard/mouse.
Unix was designed from the first to be used on a shared computer. The idea was that computers were so expensive that you could only afford one for an entire department, so you hooked up a bunch of terminals and let lots of peopleuse the machine at the same time. The "prime directive" of OS design was "it should not be possable for one user to screw up an others users work."
Unix was designed to run on very expensive shared computers while Windows was designed to run on cheap in-expensive personal computers that were owned and used by just one person. Mac OS X is based in Unix and ha very stong abillty to pertition users from each other. Untill recent years Windows did not even have to concept that there might be more then one user
I've always thought there's a slightly different phenomenon at work for Mac users.
See, Mac users really like what they're using. If you go to the trouble of buying a Mac, you're joining a group of people that is generally supportive of their computing platform.
So I think there are a lot fewer people who are really interested in breaking into Macs and damaging their computing platform's reputation.
To show this principle in action, take a look at the iPhone hacking community and how quickly they found exploits. The difference? The motivation for breaking iPhone's security was to be able to write software for the device. They were not trying to be destructive and did not see themselves as destructive.
So it would appear that there are fewer "destructive hackers" for Apple products than there are for other platforms. People are only really interested in breaking into Apple systems when there is some kind of hacking challenge, or when a product like iPhone or Apple TV is preventing them from using the devices as they wish.
I do believe that Apple has better security overall than Windows, but at the same time I also think the overall software environment is far more benign.
D
That does not mean that you can't set your kids up with a limited account--it simply means that due solely to the way Alpha Centauri is written, it won't work with a more secure setup. But that's not Microsoft's fault--complain to Firaxis.
If you haven't been down-modded lately, you aren't trying.
Sacred cows make the best hamburger.
Even if Mac OS X prompts the user before allowing a program to elevate privileges, does that matter if users just click without looking? After all, lots of programs prompt for such things in order to install some shared framework they use at the installation or first-run stage
I think in this area apple has an advantage that often goes overlooked. The number of warnings and popups a windows system presents the user with is a magnitude greater than what an OS X user sees. My god, plug in a flash drive. Can you escape with fewer than three popups? There's a reason windows users have the "click the button in the new window that just opened without reading it" mentality. Windows has gone from just letting programs do as they please, to popping up dialogs every 25 seconds. This is not an improvement, it just conditions the user to ignore the message and click the button to get on with it.
So although the windows and the mac user base will both have a degree of "make the interrupting box go away" mentality, a lot more os x users stop and read the box when it pops up, because they're not used to being harassed constantly by it and have an actual interest in seeing what it's about.
My other unrelated point is developer assumptions. I have, to date, ran into three pieces of OS X software that REQUIRE you to be logged in as an admin to either install, or to run, their software. In all other cases, they will either install or run if you provide an administrator's login and password.
And the grand majority of software for OS X does not require installation to run, you can drag it to your desktop and kick it off. Again the and also the grand majority of software for OS X, particularly that which your "average user" would want to use, does not require any authentication to run because it can function from within a basic user's privileges.
Compare that with windows. It's very hard to find a single app that will run without installing, and neigh impossible to find an app that a non admin can install. Once installed, about 8% of the programs won't run at all or won't run properly if you are not an admin user. Numerous programs will only run if you are the specific (admin) user that installed them.
Some of this is windows' fault, and some of it is programmers' fault. The programmers have come to expect all users to be admins because that's been the default. And it's easier that way, there's so much less you have to deal with if you can assume the user is an admin. So they take the short way and simply demand you be an admin to install it.
This perpetuates the problem, because now everyone wants to be an admin because they can't install software or run some software etc without logging out and back in as the admin, so for pure convenience they use an admin account.
There is a third point this just reminded me of. Assumed administrator rights. There is a group "admin" on OS X, that DOES give you write permission to certain folders that an unprivileged user does not have. But the scope is very small. On windows, merely logging in as an admin gives you a whole basket full of extra powers. This is probably why the programmers want the users to simply be admins, because it makes their jobs (particularly on installing) soooo much simpler. Instead of dealing with a dozen different permissions, you can either say are you an admin or not and be done with it. If you are, go do whatever you like. Otherwise, take a hike.
This again perpetuates windows wanting to default to admin, users wanting to default to admin, and developers wanting to default to admin. That throws a pretty daunting wrench into the works when trying to secure a system by default.
I work for the Department of Redundancy Department.
DISCLAIMER: I work at Microsoft.
Pretty much everyone who posts about this is full of shit.
Vista has had 34 vulnerabilities over the last 1.5 years. That's less than Mac OS X over the same period.
If you want to argue that Mac OS X is "more secure", you need to do it on grounds other than vulnerabilities. At best, Mac OS X and Vista are similar in the number and severity of vulnerabilities.
So the new big thing on Slashdot, since the vulnerability statistics don't back up the "more secure" argument, is to argue that Mac OS is "intrinsically" more secure than Windows.
I have no idea what people are talking about there. Vista has ACLs, just like Mac OS X. Vista has sudo (UAC in Vista), just like Mac OS X. Vista disables network-facing services by default, just like Mac OS X. Vista has a firewall, just like Mac OS X.
So, you can wave your hands and say that Mac OS X is secure because it's "UNIX". But I'm not impressed. There's nothing "intrinsically" secure about UNIX compared to any other modern OS.
What I can say is that Apple doesn't take security bugs seriously. Microsoft acknowledges when there is a reported vulnerability and reports when a fix is delivered. Apple pretends that vulnerabilities don't exist. Apple sometimes stealth-patches vulnerabilities away. And Apple frequently tries to downplay the severity of vulnerabilities.
Take, for example, the root privilege escalation vulnerability reported several days ago in Mac OS X. That kind of bug is extremely serious, yet we had 20 people on Slashdot commenting about how it's not a big deal. Apple hasn't even acknowledged that there's a problem.