Future of wireless broadband is in the air

With UK broadband take-up passing ten percent of households and BT celebrating by announcing new and ever more widespread DSL availability, the dominance of wired broadband would seem to be assured. Wireless systems have come and gone, and everyone who ran a test -- or even a production -- network has given up and gone home.

But radio will win. The only question is when and what -- in the end, not even that. Wireless is better than wired: it's quicker to deploy, costs less to maintain, has less to go wrong and is far more flexible. Wherever there's a choice between the two technologies, wireless wins. In the early days of telephones they were used to deliver music and news to subscribers, but as soon as broadcast radio came along the economics of one-to-many proved overwhelming. Open air is always cheaper than buried copper.

The lead broadband wireless technology at the moment is WiMax, 802.16's market-friendly name. Promoted at the moment as the key technology for remote, rural and otherwise unwireable locations, it promises up to 70Mbps and up to 70km range. It won't reach this in practice, but the engineering behind it is building on the enormous amount of experience the industry has from 802.11b and other wireless deployments. It'll work well enough.

802.20 is another broadband wireless standard, this time aimed primarily at mobile users. Designed to deliver around 1Mbps to devices on the move at speeds of up to 250kph, the standards committee have been looking particularly closely at the way it works with 802.11. It's a lovely idea, being able to switch from hot spot to high-speed mobile service and back again without noticing, even if nobody can quite explain why it's such a similar idea to 802.16e. That's the mobile bit of WiMax -- it uses a slightly different set of frequencies and has some slightly more restrictive speed limits, but it too is aimed at delivering broadband to the peripatetic.

The mobile phone industry is anxious not to be left out. It invented mobile data, after all, even if it's been ferociously bad at working out how to sell it or upgrade it much past the 9600bps with which GSM was born. Even though the faster data rates of GPRS, Edge and 3G networks have been hindered by indifferent coverage, the next generation is already being prepared. High Speed Downlink Packet Access -- HSPDA -- is Nokia's big idea, and is promoted as being capable of boosting 3G speeds to 10Mbps or even more. It lives alongside existing installations, just as Edge does with GSM, and is just as dependent on the networks getting it right.

All of the above can do the job of getting broadband access out to the dispossessed, and one standard will win -- most probably WiMax, with the others relegated to niche markets and more expensive roles. But it won't end there: as the market develops, costs will fall and installations will simplify. It took a couple of cycles of wireless LAN development for 802.11 to break out of its own expensive niche: as people learned how to make it work and what to use it for, it got cheaper and more desirable. The final kick up the backside was the Internet, which suddenly provided an infinite amount of things that people might actually want to transmit over their wireless network: a desire already built into wireless broadband.

Once we're into the second generation of WiMax, subscribing to the service will be as simple as buying a box, plugging it into your computer and moving an aerial around until a light comes on. You'll sign onto the service by entering your credit card number into the Web page that then appears -- just as you do with hot spots now -- and that'll be that. Deployment costs to the Internet provider? Close to zero. Equipment costs to you? Less than a mobile phone. How can wired broadband compete?

The third generation of wireless broadband will be the final integration of the telephone and data networks. You can already have a voice over IP phone that looks like a mobile phone but uses a combination of the SIP protocol and Wi-Fi to route your calls over the Net whenever it finds a hot spot. Add the mobile broadband stuff, and that phone will not only act as a mobile broadband terminal -- with phone functions hanging vestigially off it like AM radio in a £2000 digital home entertainment system -- but it will act as a local gateway across 802.11 and the forthcoming ultrawideband standards.

At some point -- and it won't be that long coming -- the economics will demand less and less human management of the system and put more and more smarts in the boxes themselves. At this point, it will become moot what standards are best or which to choose: you'll tell your machinery what you want and how much you want to pay, and it will configure itself appropriately. The technology exists: it's just a question of making it pay.

You'll notice that there's not much for the old phone companies to do here, either wired or wireless. Yes, they have worldwide networks of wireless masts -- but WiMax may be able to replicate that a lot more efficiently. And even if it can't do that from the ground, then high-altitude platforms may be able to deploy huge amounts of bandwidth across vast swathes of countryside in the time it takes to launch a balloon.

There's irony in the notion that just as the telcos are getting the hang of broadband they're in most danger of being out-evolved and relegated to the legacy room. But with the speed at which WiMax, its friends and enemies are being developed, and the clarity with which their future is being planned, the lack of compelling alternatives from the old guard promises nothing less.