Bluetooth set to take over wireless from WiFi as IEEE abandons the fight for UWB
by Guy J Kewney | posted on 26 January 2006
The IEEE has abandoned its effort to create a UWB standard, but has agreed on a draft for the next generation of WiFi, 802.11n. The conventional wisdom is that this week's events are great news for WiFi, and a disaster for Ultra-Wide Band, UWB, and by association, Bluetooth. In fact, the exact opposite is likely to be the judgement of the future.
Who was surprised when the IEEE abandoned an attempt to define the ultra-wide band wireless standard? Amazingly, lots of people. "Why haven't you covered this? It's your area of expertise!" complained a reader. "Is it news?" I asked, astonished. Apparently, it was.
But it was only news if you weren't studying UWB. Insiders knew it was coming. It's been a horrible mess for more than three years now; politics, power groups, competing technologies which are completely incompatible, and confident statements by chief technology officers which have all proved to be cobblers. And I don't mean technical cobblers: I mean, nonsense in the field of commercial reality.
The technology is not a mystery any more. There are several ways of doing UWB; the trouble is, you can't make them all part of the same standard. And the reasons aren't simple, but if you want to simplify them a bit, you could say that there are issues of power, data speed, and carrier material - because UWB isn't just a wireless technology: it can be used down fibre, too.
Ultra-wide band wireless is close to being magic. It's the ultimate expression of technology advances in the last twenty years - which is to say, it's the exact opposite of how radio has worked ever since different wavebands started being allocated to individual transmitters. Those early radios tried to reach all around the world on one frequency, so it was very important not to have two transmitters on the same frequency. So the technology originally focused on getting the band narrower and narrower.
Ultra-wide band uses the whole damn spectrum, everything from (relatively) long waves in the microwave area, right up to stuff approaching the visible spectrum. The cost is that you have to make sure you don't send the signal very far. Easy: you use virtually no power at all, and indeed, most designs of UWB are quieter than ordinary cosmic background noise.
Where UWB scores, is in sheer data speed. Because you can use the whole spread, you can shove an awful lot of bits through it. Short range, but fast. Bluetooth is also short range, mostly in order to save power. But Bluetooth gives you just a couple of megabits. In the same distance UWB could give you half a gig - which is, of course, why the Bluetooth people thing they will use UWB technology in the future.
That's going to carry on. Or, to quote Engadget's Evan Bliss, "According to Michael Foley [picture above, left] executive director of the SIG, they will work with whichever flavour of UWB triumphs in the market to ensure interoperability with Bluetooth."
But it's not market driven; the SIG will nudge things. It "will essentially pick one side over the other in helping the market determine the most advantageous solution."
Can the Special Interest Group really muster that sort of muscle? Track record: yes. The SIG has been taking a PR caning for the last four years, mostly from bleeding-edge fans who wanted to see faster, longer range technology. The SIG has simply refused to be rushed; it's taken the view that what the market wants is a stable standard, not something which makes exciting headlines for a day, but puzzles retailers for a year.
Meanwhile, what about WiFi?
The WLAN world is ecstatic: what looked like being a long, drawn-out and chaotic dispute over what the MIMO wireless standard would be, has been resolved, and 802.11n will be virtually defined by March this year.
In reality, all this has done is speeded up the trains as they head towards the site of the train wreck. Once you have a standard for MIMO, every domestic hotspot will be able to send its signal at least twice as far as before.
Why can't people see that this is retrograde? It goes against all the progress made in wireless in the last twenty years; it extends range, rather than restricting it. That means more transmitters in the same channel. And that means interference, lost data, unstable links, and data rates cut back.
Of course, individual hotspot owners want to be able to work further away from their broadband link. That's quite understandable. The whole point of wireless networking is to let you get away from the wall socket. And motor car owners want something that goes faster; the whole point of a car is to go faster than walking. But the result, in our cities, is that people are driving vehicles capable of 160 miles per hour, and averaging speeds that a bicycle can beat.
The option of moving WiFi to 5 GHz is there, and it will alleviate the traffic problems. But the option of going MIMO with 802.11a is there; the same mistake.
The solution isn't hard to discover.
If the ordinary wireless access point was capable of linking to other networks, and sharing traffic, you could access your home network from anywhere in the neighbourhood. The technology to do this - mesh technology - is tried, tested, and commercially successful. What it isn't, is standard.
So, why isn't it standard? Easy: because the IEEE mesh task group has been filibustered to death by people from large companies filling committees with engineers, and advancing the claims of obsolete technology which is proprietary to them. The attempt to achieve a standard was never sincerely made; the ideas put forward were ideas which had been shown to be either irrelevant in the market, or technically primitive. If the Intel proposals, for example, were to be adopted, wireless Mesh technology would go back five years. Of course, a cynic might say that this would suit Intel just fine, with its WiMAX ambitions, and I don't know that I could prove that cynic wrong.
It's hard to put together a convincing forecast of anything except chaos for WiFi. Not this year, not even next year; but after that, the 2.4 GHz band of 802.11b and 11g and 11n will become too congested to use. Can WiFi survive the public disillusionment that will follow? I wouldn't say so.
Meanwhile, the success of Bluetooth is likely to expand. The current spread spectrum technology at 2.4 GHz can duck ad weave around WiFi because it doesn't need to be as ambitious in terms of bit rate, and because of its inherently more efficient use of frequencies. And by the time 2008 starts up, I would bet quite a lump on seeing UWB established as a de facto Bluetooth technology, for handling the high bitrate stuff.
If the Bluetooth SIG has the foresight to build Mesh technology into its chosen version of UWB, then it will win. It won't solve all problems, because setting up mesh networks isn't just an engineering problem; there are social protocols as well as data protocols to sort out. But for those protocols to be enabled, the engineering basics have to be in place. And the challenge, there, isn't huge. There are sections of the UWB spectrum which are optimal for longer-range signalling, which can be dedicated to inter-node traffic; and others which work better for short range work, handling local devices.
What will provoke the switch to UWB and Bluetooth?
I'd say: "Laws."
Some time in the next two years, I expect to see metro area authorities start to call for the banning of WiFi except for the networks they run themselves. They'll be happy to have residents use the metro WiFi, but they won't be happy to have their own critical communications infrastructure sabotaged by seeing fifteen residents each set up a powerful MIMO device on channel 11.
Precedent says that they can do it. Several airports have banned external WiFi from their territory, despite the theory that it's licence-exempt and therefore open to anybody. If a democratically elected body bans private radio transmissions in the area, the only problem is policing it. And the operator of a city-wide wireless network will have no trouble at all in triangulating onto rogue hotspots.
That will make it necessary for people who want their own, private access to their own, private home server, to use another technology. Mesh-linked UWB networks will suit just fine: high speed - faster than you need for high definition TV, anyway! - and low powered wireless which will carry voice, data, and video without interfering with anything else.
The IEEE can't help. It has run away from the problem. The WiFi business and the WiFi Alliance are powerless to do anything except fight, and hope to keep the fighting within reasonable proportions, and reduce collateral damage. - for example, to make sure that the MIMO technology revolution isn't five competing technology revolutions.
But the Bluetooth SIG has shown that it can manage a wireless technology. As the proof-of-concept networks start to emerge, the SIG can decide which variant of UWB best suits the needs of the market, and adopt one.Of course, that product will be proprietary - but nobody in the Bluetooth market or the UWB market will turn down the option of getting the Bluetooth brand behind them - the chance of making a proprietary standard into a de facto standard without the help of Bluetooth and the SIG will be virtually nil.
And the chance of making a small fortune by sharing your intellectual property with the SIG, is pretty nearly guaranteed.
The ultimate traffic jam - You can discuss this article on our discussion board.
Bluetooth set to take over wireless from WiFi as IEEE abandons the fight for UWB