Customer Reviews

Snort Cookbook

5 star:		(1)
4 star:		(3)
3 star:		(1)
2 star:		(0)
1 star:		(0)

 

 

I read the Snort Cookbook because I am always trying to learn more about Snort. I've read almost every book on the open source intrusion detection system, so I hoped the Snort Cookbook might offer advice not found elsewhere. Unfortunately, whatever good material appears in the book is overshadowed by outdated or outright bad advice. The best Snort book is still Syngress' Snort 2.1, so I recommend reading that title.

The Snort Cookbook starts poorly with ch 1, which at 50 pages is the book's largest. After repeating installation instructions covered in online resources, the book turns to dubious packet collection recommendations. Item 1.10 suggests creating a listen-only Ethernet cable but never mentions disabling ARP traffic with ifconfig's -arp option. Item 1.11 describes how to build a homebrew tap but doesn't address signal regeneration problems that could result in traffic loss.

Item 1.12 gives terrible advice: "If your Snort machine has only one network interface, using the passive tap, run both lines to a small hub. Then from another port of the hub, run a cable to your IDS. This will combine and maybe even buffer the traffic for the IDS and give a full duplex connection." Wrong -- this is a nice way to never see traffic when full-duplex packets from the two transmit lines collide in the hub.

Item 1.14 says "Snort itself is incapable of sniffing a wireless network," but it ignores the fact that while Snort doesn't understand 802.11 traffic, the sensor can join a wireless network and interpret what it sees. Item 1.15 demonstrates more ignorance of hardware issues by saying "Linux-compatible gigabit Ethernet cards are available with up to six ports. Coupled with machines that have space for three or four PCI cards, you could have as many as 24 Ethernet ports." This suggestion completely ignores the fact that a single gigabit NIC will saturate a 32 bit, 33 MHz PCI bus, and many BIOS will not be able to handle interrupts from more than about 8 NICs in a PC.

Item 1.25 says "two to four million records is the max for MySQL," which is odd. One MySQL database I use to collect session data on Sguil has over 31 million records. Item 1.25 also covers the often-repeated and incredibly naive method of having Snort log directly to a database, without utilizing Barnyard as an intermediary. Thankfully we see Barnyard covered in ch 2, but recommended for "high-speed network[s], such as 1 Gbps or greater." Barnyard is definitely appropriate when monitoring at less than gigabit speeds.

Throughout the book, the obsolete ACID Web-based alert console appears. BASE has been available since October 2004; it addresses stale code problems in ACID and should have been covered. I was disappointed to see the Sguil suite mentioned but never given any discussion, even though the older Snort 2.1 book introduces using Sguil. Item 4.2 mentions "RST scans" even though they are a fiction of one security researcher's imagination. Item 6.6 claims to offer ways to test Snort by showing three programs (Snot, Sneeze, Stick) that have had little effect on modern Snort implementations (e.g., 2001 on).

On the positive side, in many cases the Snort Cookbook properly addresses questions which frequently appear on the snort-users mailing list. Items 2.15 and 2.16 show how to send Snort alerts to email, a pager, or cell phone using Syslog and Swatch. Item 3.2 discusses rule updates with Oinkmaster. Rule issues in ch 3 were generally helpful, like dynamic rules (3.4), evasion issues (3.10), optimization (3.13), and even Spade (3.18). Perfmon coverage in items 4.6 and 7.0 help discover how well Snort is working. I also liked the policy-based IDS ideas in item 7.5.

The back cover of the Snort Cookbook says the book "can save you countless hours of sifting through dubious online advice or wordy tutorials." That online advice is frequently more correct than what appears in this book. While some of the book is helpful, often that material has already been introduced in online documentation or best covered in Syngress' Snort 2.1. Perhaps a second edition will address the concerns in this review and produce a more useful cookbook for future readers.

The core of this book is the chapter on Rules and Signatures. Snort is renowned for its rule language and its vast flexibility. It is a reasonably high level "script" that seems more declarative than procedural. Ok, I'm speaking a little figuratively, but if you scan the rules, you might see what I mean. The chapter explains how to build rules of varying levels of complexity, depending on your needs. One neat trait is the profuse range of options for detecting traffic around the machine running Snort.

Of course and inevitably, the default rules base has grown and it is regularly updated. Currently, these defaults number some 3000, and few sysadmins have the expertise to understand all of them. So one recipe tells you how to get and run an updater program (Oinkmaster). Though you are cautioned about letting it change your rules automatically.

Other recipes expand upon the rule scope in interesting ways, like looking for p2p or Instant Messaging traffic. You might be responsible for a corporate network that bans these, perhaps. Here is a simple way to show a supervisor how you can stay on top of the problem.

Snort Cookbook O'reilly
by: Orebaugh, Biles & Babbin

What can I say designing a reliable detection system is a challenge at best.
This book makes it seem easy! I thought this was the best layout of a tech.book I have ever saw.
Problem > Solution > Discussion. they gave you the information in a precise way with out overloading you
with material you did not need. The Rules section was espcially useful...
The only downside is I wanted to see more on rules with samples.
Overall this was a very useful Book. I already had snort in place this made it much more useful.

Brett Hoff

Actually, probably everything you'd need for a tutorial is in here; it just isn't put in one place up front. Therefor, for someone totally unfamiliar with Snort, the sudden jump from installation to cook-book recipes may be confusing and unsettling.

As there is plenty of material at http://www.snort.org/docs/ and as getting Snort running isn't all that complicated anyway, that's not a major flaw.

Like another reviewer here, I think the rules sections are probably the best part of the book, though I was also impressed by the attention given to the specifics of Windows and Mac OS X - it's nice to see that level of completeness.

When the Internet was being set up, who could have possibly believed just how unfriendly a place it was going to be out there. After all, it was just a concept where scientists could exchange papers. Even if you would have told the original developers where it was going to go they would have just laughed at you.

Anyway, Snort is another tool in stopping the bad guys from coming into your system. In particular it is an intrusion detector. Note the word detector. Snort monitors your system to see what's happening. It is not an anti-virus like program that detects, quarantines, deletes, etc. an infected file. Instead it watches what is going on in the system and looks for behavior that is outside the rules.

Snort watches, records and reports on what the systems in you network might be doing. On a big network, running Snort could well be a full time job. It can produce volumes of information. Some of this information regarding your employees might be considered spying on them, there are also some words (a few more wouldn't hurt) on what you can do to outsiders vs. your own people.