For those not aware, a normal 802.11 channel is about 20 MHz wide.
The Atheros Chipset used in some of the professional grade 802.11/Wifi products (like Ubiquiti) can support half (10 MHz) and quarter rate (5 MHz) channel widths.
Obviously the maximum data rate (normally 54 Mbps for a conventional 20 MHz wide channel) drops, but even at quarter rate is still very usable with a maximum data rate of 13.5 Mbps.
The open Atheros driver talks directly to the hardware abstraction layer (HAL), and is also capable of frequencies outside of the Part 15 band.
[For a while no open source HAL's existed that can let you do 5/10 MHz mode. You had to use MikroTik, StarOS, IkarusOS, DD-WRT and a few others for these modes. As of June 2010, it appears that 5/10 Mhz support seems to be implemented in ath5k now.]
You can see the feasible channel selection overlay here:
http://www.qsl.net/kb9mwr/projects/wireless/allocations.html
The channels in light blue fall into overlapping amateur band space. And are acceptable for HSMM operation.
As you can see you can squeeze seven 5 MHz wide channels below the first Part 15 channel on 2.4 GHz, two of which are completely outside of the Part 15 overlap. (Thus, should have quieter noise floors)
On the 5 GHz band, there are thirteen channels that are completely outside of the Part 15 overlap.
And even more interesting is that that within the Atheros chip it is possible for licensed developers to enable a local oscillator generation for a direct conversion radio transceiver. This is Not an open function, but irregardless, this is how 802.11 products on 900 MHz (Ubiquiti XR9), and 3 GHz (XR3) (as well as other places) are possible and on the market.
Open source drivers unlocked the possibility of additional frequency support. It allows programmers to be able to write a driver. In summary; Atheros has allowed a third party to create a layer between the low-level functions of its chips and high-level drivers via the madwifi/ath5k development.
So what about unlocking additional channels in other chipsets/hardware?
Broadcom is the the chipset of most common Linksys WRT54G routers.
Broadcom has for a long time declined to provide non-licensed access to it's chips. A project that has been working to reverse engineer access using legal means had released its first working drivers for Broadcom 4300 series chips a few years ago.
If you are into the nitty-gritty, a recently released (2010) Broadcom wireless driver seems to have structures which imply the PHY in the chips can be directly controlled to program HSMM channels.
Also see:
http://lwn.net/Articles/456762/
http://linuxwireless.org/en/users/Drivers/brcm80211
For a slightly less technical overview of the various aspects of Modifying Consumer Off the Shelf Wireless LAN devices for specialized Amateur use, have a look here:
http://www.qsl.net/kb9mwr/projects/wireless/modify.html
I was recently asked my opinion about the use of Amateur Radio Networks like HamMesh, HamWan and BCWarn.
The question is pretty generic. It all seems to stem from the ARRL HSMM working group from a number of years ago. Despite that group falling apart, independently amateurs all over the place have embraced the technology. It's great in my opinion. As interest in voice repeaters continues to wane, multi-media networks do make perfect sense. These more modern types of networks have the potential to draw new blood into the hobby. New hams who have software skills that can help the community with software defined radio and so forth.
Outside ham radio, as consumers were now live in a world where to keep thing interesting and new we have a flexible application space. Be that apps on our phones, software on our PCs, and even firmware updates to our more hardware like devices.
That that has been notable absent in ham radio. I.e. What it is when you buy it, is what it will be 5 years from now unless you want to totally replace it for the tune of several hundred dollars.
Ham radio used to be a good starting place for many who later entered broadcast and electronics careers. Today those positions are few and far between due to disposable electronics and consolidation of engineers with mega broadcast groups. What is the most notable/abundant "tech" career today is IT (information technology) work.
In my humble and simple opinion: These types of networks are long over due, and I am glad they are continuing to grow. It helps ham radio stay relevant.
73'
Steve, KB9MWR