Scott Bradner: Will the wrong wireless succeed?

At one time it looked like there was a roadmap, a confusing one perhaps, but a roadmap nonetheless. Wireless Internet was going to be everywhere, but you were going to use different wireless technologies depending on just where you were. This original roadmap seems to be getting overtaken by events, and a far simpler one may be emerging.

The original wireless roadmap looked something like this: Local communications, say between your cell phone and palmtop, would use Bluetooth; connections to your office LAN would use 802.11; at home you would use HomeRF; and when out wandering the world you would use 3G or maybe someday 4G. Come to think of it, I never did figure out how you or your mobile computing devices would know where you were so as to know what technology to use.

Each technology has been optimized for its particular role. Bluetooth is low power and short range (30 feet or so), and slow – less than 1Mbps.

Then there’s 802.11, which offers a range of speeds from a few megabits per second to about 50Mbps at a few hundred metres. HomeRF only needs to cover a house and runs at a few megabits per second. The next-generation cell phone services, 3G and 4G, will offer a few megabits per second at distances of up to several miles.

But maybe optimization is not needed. Supporters of most of the technologies in this fuzzy picture might want to consider that “good enough” rather than “optimization” may just be taking over.

No, 802.11 is not just for office LANs anymore. It’s showing up in classrooms, hotels, airports and Starbucks. You would expect this, since these environments have basically the same requirements as office LANs. It also should be obvious that for simple Internet access-type services, 802.11 would work just fine at home. But now 802.11 is starting to show up in places that it would not seem all that well suited for.

For instance, 802.11 is starting to show up as competition to cellular-based Internet connectivity such as 3G. See for a list of San Francisco-area providers.

And with the improvements in the power efficiency of future generations of 802.11 chips, Bluetooth does not seem so important. While 802.11 is far from perfect, the current versions – 802.11b and 802.11a – have significant security and quality-of-service (QoS) issues, and having two versions could be a problem. But dual-mode chips that support both versions are now shipping and the IEEE, the developer of 802.11, is busily working on security and QoS improvements.

Indeed, 802.11 is yet another example of generalization winding up being more important than optimization. In this case, there may be rough times ahead for providers of more ideal solutions such as Bluetooth, HomeRF and, most dramatically, 3G, where the US$150 billion spent for frequency licenses may have been mostly wasted.

Bradner is a consultant with Harvard University