802.11 Wireless Networks: The Definitive Guide
Title: 802.11 Wireless Networks: The Definitive GuideAuthor: Matthew S. GastPublisher: O'Reilly & Associates Inc.ISBN: 0-596-00183-5
When its time to microwave your house, neighbors or office, this is the book you need. (Microwave as in wireless Ethernet, that is.) With some minor exceptions, this book is a great reference for all things WiFi. If you plan on deploying a wireless network or are having trouble with an existing one, most of this book is required reading--just don't read it in chapter order (more about that later). This is just the book I wished for while thumping my forehead against the wall trying to inject the same material directly from the IEEE 802.11 docs. Thank you, Mr. Gast, for saving my house while giving me a better understanding of low level 802.11.
Now, with that glowing overview fresh in your mind it seems only fair to talk about what is wrong with the book first. Don't worry, I'll get the good parts soon enough.
First off, some topics related to wireless Ethernet are not covered in this book that should be.
If you are looking for a reference to build a Pringles can/yagi, a coffee can/circular wave guide, a homemade DSS parabolic dish or other high performance, homemade antenna, you'll have to look elsewhere (like The ARRL Antenna Book 19th Edition, ISBN 0-87259-804-7). But don't look here because you won't find it.
Part 15 describes what is legal and illegal in terms of these unlicensed radio devices. This book barely mentions these regulations and incorrectly interprets some of their meaning. I'd expect "The Definitive Guide" to include part 15 in an appendix along with some helpful interpretive commentary. In particular, information on calculating ERP for directional antennas and point to point applications, where the FCC regulations make special allowances, would be helpful. Commentary on FCC certification for radio systems would be useful as well. It seems that the common practice of adding antennas that are within the gain allowed for ERP, but not specifically certified with the radio, is done often done but technically not allowed. The missing part 15 reference is a large oversight.
Virtually no coverage is given to creating "long distance" links with 802.11b. The ability to dump leased data lines with high monthly fees is an exciting and profitable application of this technology. 802.11b equipment can replace T-1 lines in some instances with virtually no recurring cost and with greater throughput (nearly 6Mbs for 802.11b versus 1.54 Mbs for T-1.) Such point-to-point links and wireless Ethernet bridging certainly should receive more coverage. The few sentences about long distance links suggest using a compass and topo map to point directional antennas across long distances; no mention is made of using a GPS to calculate bearings (to use with the compass) and a handheld mirror.
The boy scout trick of flashing a mirror to locate someone too far away to easily see really works; I've used it for long links. There was also no mention of calculating the signal loss of microwave radiation over distance, fade margins, Fresnel zones or signal polarization. All these items would be useful, particularly for long distance links. Antenna diversity was mentioned, but not the fact that both antennas should cover the same basic domain for proper operation, effectively disallowing the use of one high gain and one omni directional antenna with the thought of covering stations both close and far away. In addition, practical advice on selecting equipment housing for outdoor installations is missing. Also interesting would be information on using solar cells to power wireless relay nodes in places where AC power isn't easily available.
IP address allocation (unrouted IP space) and IP routing are not covered. For folks who aren't already accustomed to unrouted IP ranges, NAT and IP routing can become hairy when deploying networks of any size. Although not just a WiFi issue, they deserve more coverage. (Gast does get brownie points for his mention of mobile IP, though.)
A few errors, misrepresentations, type-setting problems and contradictions are present, but they are relatively scarce amongst the volumes of accurate and useful info. But in case I've piqued your interest, I'll list a few. The most notable are his recommendations concerning WEP, which are frequent and contradictory.
In the typesetting department "Re-addressing" becomes "read-dressing", which confused me for a few seconds (poor hyphenation algorithm I suppose). For the minor error department: in the WEP key length sidebar talking about the 128-bit WEP, Gast says, "After subtracting 104 bits for the shared secret component of the RC4 key, only 104 bits are secret", when he means something like "after subtracting the 24 bit Initialization Vector component of the RC4 key, only 104 bits are secret."
A little more perplexing is the nice sidebar on page 154 about "the nonexistent Microwave Absorption Peak of Water", a point he seems to ignore on page four when he says the 2.4GHz frequency was chosen for microwave ovens "because electromagnetic radiation at that frequency is particularly effective for heating water." And finally, of greater concern, he also ignores FCC part 15 regulations concerning certification when talking about the use of amplifies and specialized antennas for covering large areas--a much more serious omission.
While reading the book I sometimes found myself asking "didn't you just tell me that?" because of the redundancy, though no doubt some will find the repetition more useful than annoying.
O'Reilly maintains a web site to list book corrections, a useful service. Hopefully corrections will appear shortly on the errata site for this book.
Okay, that's most of the bad stuff. So what was good? Plenty!
Except for the omissions listed above, the book's coverage of 802.11 issues was both broad and deep. Linux installs, Windows installs, WEP, access points, network deployment, network analyzers, Medium Access Control and Apple airport issues were well covered topics. Don't let my list of what's wrong or missing lead you to believe that this is not an excellent reference to 802.11--it is. Here are some highlights.
The coverage of Linux-related issues is excellent. Gast dedicates an entire chapter to Linux and covers the various drivers and software. He lists important URLs and shows where to find these resources on the Net and gives instructions for compiling and installing them. He even includes an excellent overview of PCMCIA services in the Linux setting. A table lists the popular WiFi cards and allows easy identification of which of the two common chipsets they use. Configuration files, their location, including WEP configuration for Linux, also is covered. Gast does a great job of providing the information you need to get 802.11b running on your Linux laptop.
Gast proceeds step-by-step through an Orinoco (Lucent) 802.11b PCMCIA card install under Windows, showing and explaining the different driver options. Also in this section I finally learned what Orinoco means: "Orinoco, comes from the third largest river system in the world... 10 miles wide and more than 300 feet deep... the river's name is derived from the native words for 'a place to paddle'". So Orinoco is "a place to paddle"; I still don't know how to pronounce it.
Media Access Control (not the product by Apple, though there is an appendix on the Apple airport setup and configuration), the rules of the road for how different radios share the airwaves, is well covered in Chapter 3. Clear text, logically presented and accompanied by illustrations, does a wonderful job of covering this complicated topic, light years beyond trying to get the same information directly out of the IEEE 802.11 specs. If you need to understand the way CSMA/CA works in 802.11, this is the place to find out.
Chapter 5 describes the WEP protocol in detail and specifically discusses why it is worthless. Gast explains weak IV (Initialization Vector) keys and how they are used to recover the secret part of the WEP keys. In Chapter 16 he includes links to AirSnort and helps with compiling it, running it and using it to break WEP. There are examples of all the steps and key recovery time estimates. Gast reminds us with dry humor that "in wireless networks, the word 'broadcast' takes on an entirely new meaning." As Gast suggests, any real security should be obtained by other means, such as SSH, SSL or perhaps IPSec.
Network protocol analysis can be difficult without a good protocol 'droid, but Gast has the answer: Ethereal. One of my favorite secret weapons Ethereal has the notable tag line "sniffing the glue that holds the Internet together." C3P0 has nothing on Ethereal when it comes to interpreting network traffic. What do each of the bits in protocol headers mean and what problem to they represent? Ethereal can help you find out. Gast takes the reader though installing Ethereal under Linux and Windows. He then shows several examples of captured data streams and leads the reader through the process of interpreting them--useful for tracking down those devilish network problems that pop up from time to time. It's no failing that in the fast moving world of 802.11 the Ethereal coverage is slightly dated already; Lucent Orinoco cards are now supported instead of only Intersil-based cards as stated in the text. Overall his coverage of network analysis using Ethereal is very good.
The book recommends reading chapters 1 and 2 (the overview), then skipping (over all the protocol details) to chapter 12 and reading to the end, before returning to the middle for the gory details where necessary. I heartily agree with this order. Some of the intervening chapters even have introductions like "[This]... is one of the most difficult, most interesting and least useful chapters in the book. Feel free to skip it." Which is a good introduction, considering that the chapter covers the physical layer and how the radio bit encoding actually works. Fascinating if you are an amateur radio hobbyist (HAM) like me, though probably unnecessary for most deployments and troubleshooting. It seems a little odd, but refreshing, to have the author warn of difficult, potentially boring and less useful sections that are included for completeness.
802.11 Wireless Networks continues the O'Reilly tradition of putting animals on the cover. The colophon states that the cover animal is a horseshoe bat (Rhinolophus hipposideros) that gets its name from the "horseshoe-shaped, leaflike plate of skin around the nose." On the plus side, bats seem especially appropriate for this topic. On the minus side, I guess I'll have to stop referring to the O'Reilly sendmail reference as "the bat book". (This "bat book" is a much easier read.) Referring to it as "the eight-oh-two-dot-eleven-b book" leaves something to be desired.
Gast nimbly leads the reader through the 802.11 minefield, making the complicated concepts understandable; no easy task in a world where time units of kilo-micro seconds reign. (Only 802.11 could mangle the metric system so mercilessly.) Gast often provides key links to software and other references on the Net where appropriate. His examples are useful and numerous, with the text targeted at those with some network administration experience (particularly in an Ethernet setting). Definitely worth the read if you are doing any wireless Ethernet work, this is the best 802.11 reference I've seen. I also anxiously await the second edition, which perhaps will cover some of the omissions I noted.
Monta Elkins is a WiFi computer consultant, formerly a computer system senior engineer at Virginia Tech and system administrator for the Blacksburg Electronic Village. When not microwaving his associates, he can be reached via e-mail or on the web.