Wirelessly challenged

Intel’s alternative to Wifi solves the problems of unlicensed wireless, but will it create just as many new problems?

One of my ancestors used to make steam railway engines. You can still find some of them chuffing their way around parts of South America, with ‘Kewney’ stamped into the cast iron.

However, one of the irritating parts of the business was the need to make a completely different engine for each country, where the distance between the rails was a local standard.

So I sympathise quite a bit with Intel. Right now, it has a good deal going with the Centrino wireless brand. It adds perceived value to a mobile computer, without the cost of extra components.

This is possible because of Wifi – which is a worldwide standard. Well, not exactly a standard as there are a few types: the original 11Mbits/sec 11b standard – now beefed up to 11g and running a great deal faster; the 11n, which is just about to appear and is faster still; and the 802.11a for the 3GHz bands. These standards face serious problems in the long term.

I’ve touched on this before, but the problem with Wifi is that it uses the licence-exempt spectrum at 2.4GHz – a frequency band that anyone can use for whatever they like. Users have to restrict the power with which they broadcast with, but if you find six of your neighbours all using Channel 6, there’s absolutely nothing you can do to avoid the congestion. And it makes the wireless slow and unreliable.

When this system becomes really unmanageable, Intel hopes to have the alternative ready to roll out: Wimax. This system has been talked about for many years and is now, finally, shipping. Sort of.

It will be licensed, so we won’t have any of this ‘anybody can set one up’ rubbish and, because of Intel’s propaganda and lobbying, the same frequency will be available around the world. This means every Centrino machine will work on public wireless in any country.

For my great grandfather, of course, it really didn’t matter if a Brazilian steam engine wouldn’t fit on English rails, because no one would bring their engine over to London for a holiday. For mobile PCs, the equivalent standard matters a great deal, especially if you take your PC to a country where it’s illegal to broadcast on the only frequency that you can use at home.

OK, does that mean all is beautiful in the future Wimax garden? Not really. The problem with licensed wireless is that someone has to pay to install it. You have to buy a licence, which isn’t cheap, unless it’s for a service that won’t make any money (in which case, the licence won’t be issued). And you have to pay for people to climb lamp posts and walls and plug in the transceivers.

Proxim, which has just announced one of the cheaper Wimax solutions, reckons that in London, you’d need a lot of masts. Per square mile, probably about 60 masts, possibly more. Each unit would cost a minimum of £500, plus there’s the charge of personnel. And that’s assuming you don’t need planning permission, which is assuming a lot.

Will that work? Perhaps. It’s not mobile phone technology; it assumes that you’re sitting still most of the time. There is a mobile Wimax standard (802.16e) but it has a nasty secret, which you can find out about if you read about 802.20.

The 802.20 standard is very much like Wimax, except that it has a solution to the problem of ‘hand-off’ (what happens when a moving terminal moves out of range of one cell, into the next cell). Cellular phones do this very well. Wimax doesn’t.

So the first Wimax equipment, sadly, won’t be in western cities. It may well (believes Proxim) be used to bring cheap internet connections to remote African and Asian villages, where Wifi phones can connect – this is probably why Proxim’s solution includes Wifi as well as Wimax.

However, this leaves Intel with a real problem. Having a PC with a universal wireless standard sounds great – but one that doesn’t work in London or New York does looks like something of a problem, don’t you think?