34 of 34 people found the following review helpful
on December 1, 2006
The temptation with a massive book, such as this one, is to use it as a reference. While no doubt valuable as a quick reference for people looking to know the exact problems with any given C API ("snprintf does what differently on Windows and Unix?"), this book is best read page by page. There are surprises sprinkled throughout. Vulnerable example code is taken from real software applications, such as OpenBSD 3.6, Netscape, and OpenSSH. Of course, more than just a collection of code with mistakes highlighted, this book has a powerful methodology, complete with "Desk-checking", "Scorecards" and other useful tricks.
This book is not about binary analysis; assembly language is used only to demonstrate tricky C code.
Unlike many books with multiple authors, this is an extremely well put together book that flows naturally from chapter to chapter. The chapters on C auditing are amazing. The chapters on web assessment, while not the most in-depth chapters in the book, still contain a lot of information that is covered nowhere else (servlet race conditions, for example).
In fact, almost everything in this book is, if not new, covered more expertly than anywhere I've seen. For anyone doing software security assessment, this book is required reading. All 1200 pages of it.
28 of 30 people found the following review helpful
on November 30, 2008
This book was like a blow to the head for me. I'm not a security person, I'm not coveting ever more arcane vulnerabilities. Rather, I'm the poor guy at the other end of things: I'm a programmer. It's my job to avoid all the known and imaginable vulnerabilities while at the same time providing some useful functionality to my customers.
You bet I wouldn't like some self-styled security "researcher" tear apart my poor little programs and expose all their failings. What's troubling me, after reading this book, is that it looks very much like I hardly stand a chance. Security would be hard with the best of tools, unfortunately, at least when it comes to systems programming, the tools -- C, low-level APIs -- are dubious at best and introduce lots and lots of problems of their own. These tools hail from a happier time long ago when we were still trusting trust. I was overcome by a mixture of horror and chagrin when I saw proof in this book that not even the people writing sensitive security software (such as OpenSSH) wield these tools artfully enough to avoid vulnerabilities.
And this is where I come to the only beef I have with an otherwise comprehensive book. It's like a field guide to dangerous beasts that teaches you to recognize sabre-toothed tigers, but doesn't tell you how to get rid of them. Contrary to what the subtitle promises about preventing software vulnerabilities, there is just too little about it. This is a considerable shortcoming, in my view, as a lot of the demonstrated vulnerabilities don't have trivial remedies even after they are exposed.
Wrapping up, I feel left alone in the twilight and I think I saw a tiger over there.
16 of 17 people found the following review helpful
on December 7, 2006
As a security researcher it is often hard to recommend books to people. A lot of the things you know come from actually doing it and I find it hard to explain to people the how and the why of what I do. That problem is now solved with the publication of the Art of Software Security Assessment. The book was written by 3 people very familiar with the problems of software security and even more so 3 people who actually know what they are talking about when it comes to what a bug hunter looks for in bad code. Out of all the chapters I have read I have to be honest and say the chapter on Windows IPC (Chapter 12) is worth the price of admission alone. It describes Windows messaging and mailslots to a degree I have never seen publicly explained before but the crown jewel is the in-depth and concise explanation of the Windows implementation of RPC. A lot of the major worms you have heard of, like Blaster and Zotob are based on RPC exploits. One of the challenges for developers and security professionals alike is that RPC is generally very blackbox, meaning a developer uses a few functions to communicate but never really knows what's going on under the hood. This book solves the problem by explaining all the concepts and how to look for bad code.
I can not recommend this book enough if you are serious about security.
14 of 15 people found the following review helpful
on December 19, 2006
This should have been the third book in line to be reviewed, but as I leafed through it, once, twice, and a third time I started to realize this is one of those rare security books that has a chance to revolutionize the industry like Applied Cryptography, Snort 2.0, or Hacking Exposed. The longer you wait to read this book, the further you will fall behind. Nuff said?
Every week that goes by we see an increasing understanding in the community about how important secure software is and that it takes the appropriate development process to create secure software. This book is hitting the marketplace at the perfect time, I hope the authors and publishing team have a runaway success, you deserve it. I also hope people will be encouraged by this book, secure software development is certainly possible, this book clearly shows that. It takes management support in terms of resources, training and good process, but it can certainly be done.
At 1128 content pages, much of this material will be things that you have picked up in other places, such as other books or courses you have taken. Much of it will be things you once knew and forgot. But this is the most complete book on software security out there covering Windows, Unix, Network Protocols, Web and other Applications.
What I particularly love is the how approachable the majority of the information is. Please do not get me wrong, if you have never written a line of code you are going to be lost during the code examples, the only signpost you get is the occasional bolded line, but you will still be able to clearly follow the discussion before the code example and right after the code example.
Section 1 of the book is called an Introduction to Software Security Assessment. I was able to read that 164 pages all at one time ( though I was up to 2 AM doing it). This is foundational material and if you are responsible for software development as a manager, I recommend you read at least this one section.
The next section, Software Vulnerabilities, starts with a buffer overflow chapter. This is a test of any good security book. If they point to an ancient paper like Smashing the Stack and mumble an incoherent sentence or two, you know they probably don't know what they are talking about. This book builds the case, uses both code fragments and clear diagrams with plenty of explanations.
The final section is titled, Software Vulnerabilities in Practice, I am not convinced this is an accurate section name. Network or Web should probably be in the name, chapters include Network Protocols, Firewalls ( probably the weakest chapter in the book), Network Application Protocols, Web Applications and Web Technologies.
6 of 6 people found the following review helpful
The Art of Software Security Assessment: Identifying and Preventing Software Vulnerabilities by Mark Dowd, John McDonald, and Justin Schuh is a comprehensive treatment of how to study and assess the security of your software and operating system platforms. If you are at all responsible for the security of software in your organization (and that includes all the developers, too), then this is a book that should be resident on your shelf.
Part 1 - Introduction to Software Security Assessment: Software Vulnerability Fundamentals; Design Review; Operational Review; Application Review Process
Part 2 - Software Vulnerabilities: Memory Corruption; C Language Issues; Program Building Blocks; Strings and Metacharacters; Unix 1 - Privileges and Files; Unix 2 - Processes; Windows 1 - Objects and the File System; Windows 2 - Interprocess Communications; Synchronization and State
Part 3 - Software Vulnerabilities in Practice: Network Protocols; Firewalls; Network Application Protocols; Web Applications; Web Technologies
Rather than just dive right in to detailed hacks, the authors take a measured, structured approach to assessing the security of software. The first part of the book covers the general process of reviewing for security, including design security, operational security, and application security. They also present the general areas of potential weaknesses that you need to look for in each development stage. Instead of just saying "look for bugs", they present different approaches to reviews that each have their strengths and weaknesses. You come away from Part 1 with a practical methodology that you can use immediately to consistently review all parts of your development process. Parts 2 and 3 are a bit more like other security books you've possibly seen, but much more emphasis is placed on understanding the "why" behind the problem rather than just the "how" of fixing it. Armed with this deeper understanding of why certain techniques are lacking, it's easier to change fundamental coding habits rather than just fixing problems as they're discovered in testing (or unfortunately in production). Many of the examples are in C/C++, so if that's your language of choice you'll get a lot more out of the book than others. Still, a competent developer should be able to follow the concepts regardless of their language of choice. And it really doesn't matter if you're just Unix or just Windows. Both sides are covered...
This is definitely not a small book (close to 1200 pages), but it's not padded or fluffed out to get there. It delivers real value for your money...
10 of 12 people found the following review helpful
on December 13, 2006
This is a well written and authoritative book that looks at security primarily from the perspective of a code auditor but it is also of great value to software developers. The book provides instructive background material in relevant technical areas as well as including many, many examples of actual vulnerabilities. These examples are distilled down in a very skilled manner so that the underlying coding flaw and the resulting vulnerabilities are apparent. One way they do this is simply by using bold text to highlight the important parts of the examples, which saves considerable time in reading and comprehension. The book contains examples from numerous languages but the majority of examples are from C. The book also contains "tracks" for UNIX, Windows, Web, and Network applications from which the majority of examples are drawn. The book is representative of a new trend in software security books that are well-grounded in the underlying disciplines of computer science and software engineering.
5 of 6 people found the following review helpful
on February 22, 2007
The long awaited revolution - I kid you not.
This is an amazing, comprehensive, and well-organized compendium of unprecedented scope. Yes, it has everything - all done well. If you are into security testing and live by it every day, you are still bound to learn a lot, to re-evaluate the things you know, and to genuinely improve your results. If you are a software engineer, it *will* help you build superior applications. If you are just an enthusiast, you will genuinely enjoy the time spent with this book, and you will find this brick handy more often than previously imagined.
Now click "add to cart". Really.
Dowd et al have produced a vast excursion into how to analyse software for possible security flaws. This book was published around the same time as another related text, "The Art of Software Security Testing" by Wysopal et al. The latter is much shorter and offers a quick overview. Dowd's book is far more indepth. It gives more examples of many ways that code can inadvertently expose a computer to attacks.
There are numerous general guidelines. Like fixing simple bugs first. Partly so that you can demonstrate some tangible progress, especially if you have been called in as a consultant on someone else's project. Another tip is that automated source analysis tools are often weak.
A lot of the text centres on code written in C. Many bugs are ever-present. A big cause is the weak string handling routines, that often do not test for buffer overflows. Pages and pages go on at considerable length about this. It is not reassuring to read that C++ has also inherited these weaknesses from C. The narrative explains that part of the attraction of more recent languages like Java or C# has been their more robust string classes. Along with automatic memory allocation and deallocation.
A commendable aspect of the book is its evenhanded treatment of bugs associated with Microsoft's operating systems and those in the various unixes and linux. You get to appreciate that no modern operating system is free of these.
Despite the book's length, it cannot cover all the known attacks, let alone anticipate future ones. The authors encourage you to take the ideas given in the book as background, and possibly as inspiration in deducing future attacks.
1 of 1 people found the following review helpful
on May 6, 2014
Format: PaperbackVerified Purchase
Quite a book. Not much on web application specifics, but much heavier on C, UNIX, and security of software. It will however serve as a handy reference in the future, as most of this information will not go out of date.
on August 28, 2013
Format: Kindle EditionVerified Purchase
I bought the Kindle version of this so that I could read on-the-go, and I have to say that in some places the formatting makes it hard to follow in the examples. The hard copy is much better, although significantly less portable. I'd recommend the hard copy, given the choice between the two.