News

Intel leans heavily on wireless world with UWB, Wimax plans

Don't ask whether Intel can succeed in imposing its preferred standards, for ultra-wide band wireless, or for 802.16a local wireless broadband. Study your history. You'll see that it can.

Next time you take your computer (with built-in modem) on the road, consider what a trivial matter it is to connect it to a phone network - anywhere in the world. All you need is a cable with the right attachment on the end. And that's possible because Intel leant, heavily, on the approvals people.

In the early days of email, it was actually illicit to plug a modem that was designed for use in America, into any European phone network. Heck; it was illegal to plug a French-approved modem into the German network, or a Spanish-legal device into a UK phone socket.

Intel stepped in, and put a modem on the market. It included a small flash memory, into which, Intel said, the configuration details for all countries could be loaded. It then submitted this device for approval into most major countries. And, one by one, the resistance of the authorities crumbled, and approval was given.

Intel then withdrew from the modem market, and proceeded to sell flash memory to every modem manufacturer. (It now dominates the flash memory business).

What Intel did in the late 80s and early 90s is now being reflected in wireless. At least, that's clearly the plan.

Sean Maloney, who runs Intel's Communications Group, is committed to making all sorts of wireless ubiquitous.

His reasons aren't even vaguely secret. Intel (as he has said) believes that in another silicon generation, most radio devices will be built in Intel's favourite semiconductor technology: CMOS. Intel reckons it will have the best CMOS wireless chips; it wants to set standards that make these chips sellable.

The threat to other chip suppliers is simple enough. No, Intel doesn't want to enter the market for making WiFi and GSM and CDMA and other wireless chips. Instead, it wants to build these devices as standard peripherals on its processors.

So, in the year 2009, you'll go out and buy your 128-bit Intel based wrist-watch and built into its high-speed arithmetic/logic unit there will be a universal, configurable radio. Software will drive it; it will switch between all different standards.

In other words, it will be able to function as a GSM phone, a UMTS radio, a WiFi device, and even a UHF radio or TV receiver - or transmitter - depending on what your program requires.

The important point is: Intel will build it all on chip.

This strategy, again, is hardly news. "Native Signal Processing" meant that the Pentium replaced complex (and costly) digital signal processing chips in commercial modems. Software controlled what signals to send down the phone line; modems became much cheaper. Good news! - except, of course, that if you wanted to run something other than Windows, your software ceased to work, and your modem ceased to function.

We're seeing something similar with WiFi, where your notebook PC will drive almost any sort of client adapter card, and search for a local hotspot - unless, of course, you're trying to run Linux or FreeBSD. It's a driver issue, not an integration issue - at the moment.

But the Centrino process shows what Intel wants. A Centrino notebook has a built-in WiFi wireless chip set. If you buy a Centrino notebook, the only reason you could have for ever buying a WiFi card, would be that Intel made an error of judgement, and didn't make the right standard Centrino wireless.

That's a constant problem for Intel and Centrino, as long as it has to integrate wireless chips made from ordinary analogue technology into the motherboard. What it wants, is a motherboard which can be updated, just by downloading new software - a "soft wireless" technology.

Hence the standards battle. Intel has no use for an ultra-wide band wireless standard which it reckons can't be supported in a CMOS-based, programmable radio - eventually. Motorola, on the other hand, will have very little market for its radio devices if everything you can buy has a Intel processor with a built-in Intel soft wireless already working.

For the WiMAX forum, having Intel come on board with its support and influence will seem like a big win. This will make it happen! - well, yes, it will. But what, exactly, will Intel want to see happen? It's easy enough: it will want the standard to evolve in a direction that suits its software controllable wireless plans.

Resistance, as they say, is futile. If you're working in wireless, you probably will be assimilated - this is The Incredible Hulk Of Borg.

Well, all sorts of possibilities exist in the future. But Intel has made its intentions perfectly clear, and that is the future it has planned. Ignore the warning at your peril.