Intel poised to roll out 54MBps WLANs in Europe
I have a cunning 802.11a-ready plan...
Intel is set to roll out high speed wireless LAN kit based on the 802.11a standard in Europe from the start of next year. Which is a pretty good trick, given that 802.11a is currently illegal this side of the pond.
But before you reach for your regulators, we should assure you that the would-be Wavezilla will be shipping dual standard wireless access points with an 802.11a-shaped hole in them, pending what it sees as the standard's inevitable approval in Europe. The units have two slots which will take cards "close to mini-PCI," so initially they'll sell in Europe with one filled with an 802.11b card, and as and when central and country approvals are granted (which won't be all at once) you can just go out and buy the 802.11a version.
802.11a increases throughput to 54Mbps, as opposed to 802.11b's 11Mbps; 802.11a is - possibly - close enough on the horizon in Europe for it to stall the rollout of wireless networks, so you can see why Intel's cunning plan of selling it while not selling it has been devised. Intel is meanwhile sure that most of the people and organisations who count in Europe are already gagging for 802.11a, and that the necessary deals will be/habve already been cut to allow its sale. The unfortunate death of obscure and largely unloved local contender Hiperlan will be collateral damage.
According to Intel EMEA head of Wireless LAN product marketing David Bradshaw, ETSI is pro approval, as are numerous local regulators. The UK Radio Authority is the most pro in Europe, he says, and Brussels seems to be onside too. He postulates two possible routes to approval for 802.11a - a cunning ETSI rule change that would call it a "type of Hiperlan," which would mean it would technically have been legalised already, or a worst possible case scenario where ETSI would make a formal rule change. That would kick back approval until Q4. Once it is approved centrally, it will then be up to the individual countries to approve it locally.
Opposition to approval for 802.11a, which occupies the 5GHz part of the spectrum, comes largely from the military and satellite operators. Bradshaw is bashful about precisely which satellite operators, but it's been alleged to us that Globalstar and friends for some bizarre reason want to retain the transcendental state of peace their efforts have achieved there. Keep the chatter down boys, we're trying to sleep...
The technical objection to 802.11a is that it does not include Dynamic Frequency Selection, which allows devices to change channels to avoid interference, and Transmission Power Control, which reduces power as a device gets closer to a base station. Both of these are mandatory in Europe, and are being incorporated in 802.11a. As far as we can make out 802.11a users in the States just get to happily cook the immediate locality, or something.
Bradshaw also provides tantalising clues to some shifty business that might or might not have gone on in the backstabbing standards-setting bodies. 802.11g was originally supposed to double the throughput of 802.11b while continuing to occupy the same 2.4GHz part of the spectrum. But 802.11g is now seriously late, and has been beefed-up to hit 54Mbps. The reason for the delay, claims Bradshaw, is that Intersil and TI were wrangling like crazy over whose silicon it would use.
Now, the happy outcome of this, from Wavezilla's point of view, is that 802.11g will now roll out after 802.11a, whereas it was originally envisaged as an upgrade path from 802.11b, with 802.11a to follow. Cunning stunts like Intel's dual standard access point however present 802.11a as the upgrade path from 802.11b, so by the time 802.11g gets here, we'll all be saying, well, what's the point? And guess whose silicon...?
Icing on the cake is possibly the likelihood that 802.11g will arrive just as Bluetooth finally starts to take off, so the kit that does get out there will be done to death by Bluetooth chatter in the same part of the spectrum. Intel of course didn't plan any of this, it was all TI's and Intersil's fault, honest. ®
Sponsored: Hyper-scale data management