Slashdot Mirror
Jean Tourrilhes On Linux Wireless LAN
mcleodnine writes "Jean Tourrilhes of the Linux Wireless LAN Howto project took some time to answer a few questions from members at LinuxQuestions.org. Among some of the more interesting commments was his pick of best and worst Open Source friendly vendors ('Some of those TI engineers even sent me e-mails criticising some features of the Wireless Extensions'), an opinion or two about the Next Big Thing in wireless (MIMO), and a poke in the eye for OS zealots of any religion."
There is a Mesh AP project that has VoIP support.
More info here.
http://www.locustworld.com/
-Steve
Take a look at the LQ HCL: http://www.linuxquestions.org/hcl/index.php
I feel your pain. The problem you will run into is that even if you get a D-Link, Linksys, or Netgear, some of them may use the same chipset. I tried the D-Link first because it was cheap. It used the Prism2 chipset. But after reading every bit of documentation, and trying it as both a module and in the kernel under about 10 configuations I gave up. It seems like getting these cards to work is hit or miss for a lot of people based on all the posts I read. And believe me I read a bunch.
Then I tried a Linksys that I use on my work XP laptop. No dice there. I forget the chipset on that one, but I had the Linksys WPC11 v4 which apparently has little or no support on Linux.
So finally I decided to just bite the bullet and get a Cisco Aironet 350. If you buy these new, they are over $100. But if you go to Ebay, you can get one for around $50. All you do is compile the support in the kernel and it works like a champ. I have set it up successfully on both Debian and Gentoo.
So Cisco is the easy way to go if you can get a good deal. I would avoid the Linksys card I tried, but apparently versions prior to v4 work better. You can give the D-Link or Netgear a shot with the Prism2 chipset, but you may have to work at it a while to get it working.
Hope that helps.
As far as headaches, I think you'll find more headaches in the peripheral support infrastructure than in the wireless hardware and drivers. If you are going to use PCMCIA/PC Card wireless adapters I think you'll discover the Linux PCMCIA drivers have a habit of panicking. With any hardware you'll need to do a lot of manual configuration hacking before your computer will perform useful functions like automatically roaming to available SSIDs (something windows and mac os do automatically). You'll be installing packages and editing /etc files for the next month, but eventually you'll have something that works 62% of the time.
As hinted at in this interview, the future of wireless networking revolves around many variations on a few core themes: diversity, versatility, and scalability.
Diversity is accomplished through MIMO and other technologies like beam steering to provide a robust communication channel between wireless devices.
Versatility comes with open source firmware / drivers and software defined radios. There is no way manufacturers can foretell all of the desirable uses and functionality consumers want in their products. The most useful systems will be those that are versatile and can adapt to new protocols, encodings, etc.
Scalability can be achieved through robust ad-hoc routing protocols and decentralized security methods to produce a system that scales easily as participating nodes join and part the network without complicated provisioning or a reliance on centralized and limited backhaul or access point functionality.
There is still a lot of interesting work to be done in these areas, but the real fun starts in the applications that will utilize these new ad-hoc networking infrastructures.