S E R V E R   S I D E
View current page
...more recent posts

I know just little enough to still hope for a really cool big breakthrough: unlimited decentralized wireless bandwidth everywhere. Or at least everywhere there are other people.

Ad-hoc mesh networking, in other words. Taking the telecom infrastructure that currently underlies the internet and distributing it out into end user devices. In this new model you don't just connect to the internet, you are the internet. Instead of your modem connecting to a router (say, at your ISP) your modem will also be a router. And so will everyone else's.

O.K., that's probably a bit too simple, but something like this is happening. And the key will be software, not hardware. Specifically, software defined radio, like gnu radio. This type of software allows your general purpose computer to reconfigure itself on the fly to work with any known radio format.

Here are Cory Doctorow's short notes from a talk by the gnu radio guys at ETcon.

GNU Radio is a free software toolkit for realtime signal processing things -- radio included. Works for sonar, medical imaging, etc.

Get as much stuff as we can into software, out of hardware.

Turn all the hardware problems into software problems -- all wave forms are encoded, decoded, modulated and demodded in software.
This flexibility will unleash a tidal wave of experimentation. Think how hard, and expensive, it is to introduce a new wireless technology. The telcoms have to roll out all new equipment (like build towers everywhere!) while the public has to buy all new wireless cards. This sort of outlay cannot happen very often, which is why we still have very poor wireless technology deployed.

But move all the custom hardware into software, and now we can reconfigure as easily as we download patches for our software, or firmware updates for our machines. This is made possible, generally, by the incredible computational power of ordinary general purpose computing. Or, in other words, the need for building custom (expensive!) hardware to solve technical problems decreases as general purpose hardware increases in power.
What can you do with it?

* Conventional radio

* Spectrum monitoring

* Multichannel -- one app sucks down the whole RF spectrum; AT&T could support all its legacy phones (GSM, CDMA, Analog) on one tower, without any forklift upgrade.

* Morph mode

* Morph on the fly -- a device that reconfigures itself for what you need, sat phone, cell phone, pager, etc, your 802.11b could talk 802.11g once it's invented

* Better spectrum utilization -- listen-before-talk, then choose an unused band. Accommodates legacy users and lets you move forward.

* Cognitive radio -- minimal power, shaped xmission, etc
And this is all in software, on the computer you are using right now. Go gnu radio!

- jim 4-24-2003 6:47 pm [link] [1 comment]

older posts...