Slashdot Mirror
Linux Unwired
Infrared, Bluetooth, 802.11 (in current a, b and g offerings, 802.11i is also being discussed), wireless access points friendly to Linux, United States commercial cellular networks and GPS systems are all covered in a single title that surprisingly fits all of this information into 284 pages. All the chapters can be subdivided into two large parts - familiarizing yourself with the technology (the primer on GPS is pretty good) and running Linux on it (with code and shell command samples and lots of URLs).
Introduction
The authors start up with introduction to wireless, intended for Linux geeks who are not quite up to speed on radio technologies. The concepts of waves, spectrum and radio wave behavior are explained, so later the reader can explain what a retracted radio wave is. Then the first chapter moves on to explain antenna behavior, wireless infrastructure modes and some common problem, like a hidden node in ad-hoc infrastructure. The chapter is well-written, and you're not expected to have an advanced radio degree or ARRL membership to understand the terms.
WiFi cards
Chapters 2, 3 and 4 deal with connecting a Linux desktop or notebook to a wireless 802.11 network. The first issue is that of chipsets used in the wireless card, and even though enough research has been done already, authors discuss different quirks relevant to Intersil Prism, Lucent WavelLan/Orinoco, Aironet/Cisco, Symbol, Atmel, Atheros and Broadcom chipsets. We need to discuss chipsets instead of discussing the actual wireless cards, since some hardware may be shipped under the same brand name with different internals. "A good case in point: the D-Link DWL-650. This radio card initially shipped with a Prism II chipset and was very popular, because it worked on a Linux box. However, D-Link changed chipsets when it released the DWL-650 Version 2, choosing the ADMtek chipset. It is very difficult to tell from the packaging which version of the DWL-650 you are purchasing".
The chapters are done in traditional walk-through mode. They are not HOWTOs or compendia of reference information, available from the manufacturer's Web sites. The authors made an effort to ensure the reader is capable of starting up a wireless connection on Linux box, knowing nothing about it while learning important technology in the process. Certain wireless drivers need to be compiled into Linux kernel, so the task is not for the meek, but with detailed explanation, plenty of URLs and nice fonts and paragraph formatting O'Reilly Publishing uses to differentiate between the text, commands entered at the shell, and URLs, the book is easy to read.
Chapter 3 (available in PDF) teaches the reader how to connect to existing wireless network once the wireless card has been recognized by the system and proven functional. By the time the book hit the stores it was already a bit out of date, since the very first hotspot operator, Cometa Networks, shut down in May 2004. Chapter 4 discusses wireless security, touching WEP settings, a $20 Linuxant utility allowing the user to implement WiFi Protected Access, as well as authentication utilities wpa_supplicant and XSupplicant.
WiFi access points
The issue of WiFi access points is not trivial either, as many vendors out there will ship the product with a Windows app being the only way to set it up. However, for the access point setups that are Web-based, a browser in Linux will do the job. The most Linux-friendly access point include Linksys, Netgear, D-Link, Cisco, SMC, EnGenius, Belkin, US Robotics, Microsoft and ActionTec. Again, harsh reality kicked in between the time the book was written and went to press, and it's sad to see yet another Linux-friendly access point vendor quitting the market.
Not satisfied with commercial offerings out there? Chapter 6 takes the reader into the task of building your own access point. Don't forget that an access point doesn't need to be a compact portable - your old 486 with Linux on it and a wireless card connected to it might serve the purpose. Unfortunately, after all the hardware is bought and assembled, the final product might still cost you the quadruple (in case you go with smaller form-factor motherboards and CompactFlash cards for software storage), so consider this more as a geek project, not a viable solution. The authors use LinuxAP distribution for this task.
16 pages are dedicated to hacking Linksys WRT54G access point with Sveasoft, described as disruptive technology by Robert X. Cringely. The authors also take a brief look at Wifi-box and OpenWRT.
Other wireless technologies
Bluetooth, Infrared, cellular and GPS chapters follow the same chapter plan - first the basics of the technology and simple use case scenarios of what you might use it for, then the hardware needed to implement the wireless technology, available Linux software to do the jobs, accompanied with the list of shell commands to successfully talk to a wireless product, and after that typical applications of the working link.
The authors tested various wireless data plans in the United States, although this data, once again, is constantly changing as the operators buy one another and introduce new data plans. The winner of the quality and the fastest download tests, by the way, was a Motorola v120e phone on Verizon Wireless network. In upload speed tests a Merlin C201 PCMCIA card on Sprint PCS network won. T-Mobile also offers a PCMCIA card for its GPRS network, so the authors install and run it under Linux in Chapter 9.
The last chapter discusses using Linux computers with GPS devices and open-source GPSdrive project for reading GPS data.
The book
For those just venturing into the wireless world, the book would be useful. All the information provided on WiFi connectivity can perhaps be googled and found in various HOWTOs. With wireless operators, GPS systems and Infrared connection one would have to rely on enthusiast sites and newsgroups. Having such informative title that covers all of the technologies would be very useful to a Linux enthusiast.
With that, the book can be quite overwhelming, although it's probably not intended to be read from page 1 to the end. There's usually more than one correct way to do things in Linux, and for each successful project another competitor appears on SourceForge the next day. I like the authors' approach of dedicating most of the chapter space to one, leading, Linux package that seems to be dominant in the field, and then briefly mentioning the others. A notable omission is Intel's Centrino drivers for Linux, as the company is bound to become a leader in the chipset marketplace with 42% of notebooks shipped in 2003 running Centrino chipsets.
But overall the book proved to have a high informational and educational value, not only you follow the steps on setting up wireless technologies on Linux, but you also learn the internals of the technology and why certain things are done that way, but not another.
You can purchase Linux Unwired from bn.com. Slashdot welcomes readers' book reviews -- to see your own review here, carefully read the book review guidelines, then visit the submission page.
From the article --
Microsoft disputes the notion that there's a problem with the way Windows XP works with Wi-Fi.
That's a blatant lie, typical Microsoft attitude.
For the longest time, the wi-fi connection of my notebook would keep dropping and I thought it was because of a bad wireless card.
I changed my wireless card, I tried everything possible.
What I had not noticed was everything I was in Linux, this never happened - no matter what! There are areas in my school where the wi-fi signal strength is particularly weak, and even in those areas I never lost connection from Linux.
Somehow, when I would boot into Windows, my wi-fi link would keep dropping. I still haven't figured the problem yet, I just use Linux instead everytime.
And so contrary to what Microsoft may say, there is a problem with Windows XP (I have the problem whether am on XP Home or XP Pro). I wonder whether they EVER admit their mistakes.
The problem is if Windows can find a network that broadcasts it's SSID when yours doesn't, it will try to switch. There are three (or four) soltuions.
I REALLY hope that they fix this in SP2, because it's my number one complaint.
Comment forecast: Bits of genius surrounded by a sea of mediocrity.
Don't get me wrong, I'm not a luddite and I think these are great technologies, but for myself they just seem expensive and not terribly useful. What I do find interesting is use of WiFi for rural broadband. But that's still pretty novel.
I was one of those users who read about DWL-650, went to buy one and came back with a DWL-650+ without noticing it. It sucks. At my Computer Sciense Departament (like 12 AP), my laptop gets frozen because the driver and the computers turns very unstable. At home it works well because I have only one AP. Wireless tools are very primitive also. You can't scan networks not being root. I started writing a KDE tool emulating the funcionality of the new Win XP Service Pack 2 wireless tool but libiw is a pain. I ended importing a patched sources (from some Ximian guy) to use it. Still havent figured how to scan available networks being a user. Design flaws from a server oriented operating system.
After going through many different drivers and kernel compiles, I Later found that some 650's not only have an ADMTek chipset but some also have be a Realtek chipset. After trying the ADMTek drivers I found my particular card had a Realtek chipset and it came right up using the ndiswrapper driver using the windows Realtek driver (the driver supplied by d-link on the cd would not work. I had to get the windows driver from Realteks site). The thing works like a champ now with Fedora core 1.
I am an avid unix/linux user and I pretty much prefer to run my boxen in non-Windows mode. I also am an information security professional and use many tools such as Kismet to do wireless discovery. After much reading and research I decided on a Cisco/Aeronet card as my card of choice for both sniffing and wireless use in Linux. Well low and behold once I got the Cisco card, which was stated as being supported in the OS, I managed to get it installed and compiled to be used as a sniffer but to date and after a solid year of trying I have not been able to get the damn thing to work as just a network card. After months of frustration I finally gave up and went and bought an Orinoco card that I thought I would use for both sniffing and network access. Again, I managed to get things compiled and working this time with network support for the Orinoco card and no ability to sniff. Although both vendors claim full functionality within Linux I am to date still carrying around two wireless cards to get the job done. Sadly, when I boot Windows and plug in either card in XP they both seem to just work. Ah, the bitter irony.
as much as the many complaints about xp & wifi I have not had as many problems, I have found if you turn off wep (which is pretty much useless anyways) and leave only mac address protection to your wireless router the amount of lost packets/connection drops to next to nothing. Mind you it does mean someone might find out im posting at slashdot (oh the horror!) and I have to manually encrypt files that have sensative material (which you should anyways) Under Linux though as much as once the connection is started its great, getting that 1st connection is hard as hell, showing windows (in general) one point where it beats linux consistantly (and its not our exactly our fault either) hardware support.
I didn't read the actual review. I fear that I am already perminantly brain damaged from the summary.
I swear it said something like "This book is not about what the title says it is about, except that three chapters actually are about that, but the rest isn't. But that stuff is related. Fuck Mocrosoft."
Oh, and somehow it lead into all that with a reference to sock puppets.
I might be wrong, but I am mortally afraid of re-reading the summary given the results of the first read.
Is the actual review any better?
-Peter