The Truth About WiMax

3:30 PM -- Here is a good article from Dave Farber's IPer list that cuts through all the hype about WiMax. Point-to-point wireless will have a role to play, but I don't suspect it will ever be an equal competitor to wireline services for high-speed broadband, except in niche markets. As the old joke goes, wireless technology will catch up to wireline capabilities in two years time; and this statement will always be true.

This comment comes from reader Robert Berger:

WiMax ranges much as ten miles, while the WiFi we know today reaches only a few hundred feet. Intel wants to use WiMax to allow someone to get wireless access on a laptop anywhere. We all should know by now that the answer to all questions of RF coverage and capacity is "it depends."

WiMax has no magic over WiFi, just additional cost and extra applications of hype. The modulations are basically the same. The WiMax MAC has no more than a 4 dB advantage over WiFi when you compare them at the same power levels and antenna gain. That is dwarfed by all other considerations, such as the fact that multi-GHz frequencies barely propagate thru any physical object – including trees, especially wet trees. Trees are about 15-20 dB of attenuation; most household walls are 10-15 dB.

Licensed WiMax is allowed to have more power than WiFi. Current WiMax base stations operate at around 34 dBm of output power; commercial muni-wireless WiFi APs operate at around 24-26 dBm. So the difference between Licensed WiMax and unlicensed WiFi transmit power is about the same as one tree of obstruction attenuation.

To get to a "laptop anywhere," the laptop has to have a transmit power that is at least in the same ballpark as the transmitter. Otherwise the laptop can hear the AP, but the AP can't hear the laptop. A laptop has severe power constraints, so today most laptop WiFi transmit power is around 13-18 dB. Its a low probability that laptop vendors are going to be putting 24 dBm radios into laptops.

WiMax CPE (subscriber units) that are being sold in a fixed configuation have 20-26 dBm of transmit power. As do the WiFi CPE that are being sold for muni wireless (such as PePLink Surf) and are also less expensive than the WiMax CPE.

This asymmetrical transmit power issue is currently a problem for muni WiFi deployments, as are building and tree obstructions. It will be no different for WiMax. But WiFi APs are so dramatically less expensive than WiMax that you can afford to have many more of them, thus creating more coverage at a similar cost. And as I have been saying all along, WiFi is recapitulating the evolution of Ethernet, while WiMax is following in the footsteps of all the failed competitors of Ethernet.

That does not mean that there will not be business successes in WiMax, just as there were some VERY successful ATM and Token Ring companies (for a while). Intel and others are pouring capital down the WiMax drain, so it will be good for those companies that can pick off that capital while its flowing.

Maybe Beceem is one of those. Turns out, Intel is having to resell Beceem's chip for Intel's WiMax efforts – even though Intel is spending $1 billion on distributing WiMax. You'd think Intel, the chip giant, would want to sell its own chips. You look behind the curtains, and there's only this tiny startup Beceem doing it all. Intel recently sold off a mobile handset chip unit too, so it doesn't have too much momentum in the mobile space. I wonder how much longer Intel will continue to pour the billions into WiMax after they have divested themselves of it?

— Bill St. Arnaud is Senior Director of Advanced Networks at Canarie Inc.