What’s next for Wi-Fi?

Expect reliable multimedia capability early next year

In the home network industry, the next leg of the race is all about providing a reliable transport for multimedia — video and voice applications — an alternative to wired Ethernet. What’s at issue is the ability to provide quality of service (QoS) so those applications run reliably, with no jitter, frozen frames, echoes or dropped calls.

Back in February, the HomePlug Powerline Alliance announced the completion of the market requirements for its HomePlug AV spec for multimedia. But that really just kicked off a long process that won’t yield HomePlug AV products for nine to 14 months, the group says. Then in June, the Home Phoneline Networking Alliance announced it had completed its multimedia spec. And while a worthy accomplishment, not too many of us are jazzed up about running video and voice over phone lines. What we really want is to do is run multimedia over wireless — for the IEEE 802.11e Working Group to complete its QoS specification for 802.11b, 802.11a and 802.11g technologies. 

The 802.11e spec has been in the works for a few years, slowed by a religious war between two camps that want to provide QoS different ways. Although the lines aren’t drawn evenly, one camp roughly encompasses consumer electronics companies that want to stream video; the other is focused on providing voice. With the technology best suited to video, called “polled access,” the access point schedules time individually to each device on the network, based on its needs or required bandwidth.

With the other, called “enhanced distributed coordination function” (EDCF) data packets coming onto the network are tagged with a certain priority based on their purpose. An e-mail packet will have a low priority, a voice packet a higher one, a video packet a still higher one. The timers dole out numbers to packets based on function. Packets requiring very high availability, like video, get a lower number, which ensures they’ll have more frequent access to the network. Packets for Web browsing will get a higher number, which ensures they’ll wait a lot longer for a turn.

The bottom line is that the polled access method gives the data a 100% guarantee of service, but it’s quite difficult to do. EDCF provides about a 95% guarantee of service and is simpler to do. The video folks favor polled access; the voice folks favor EDCF. That’s what all the fuss has been about.

The good news is that a compromise has finally been reached. Wi-Fi Alliance President Dennis Eaton says the 802.11e spec should be completed in January or February 2004. (I could argue these folks knew the market wasn’t yet ready for multimedia, allowing them to argue as long as it was practical.)

The compromise involves a hybrid of polled access and EDCF, a solution that could end up causing confusion among users. 

Eaton says for vendors to comply with the spec, they must include EDCF, but parts of the polled access technology are optional. Optional means a device on the network must be able to respond to a request for polled access capability, but they don’t have to support it. What that means in real world usage who can say? Will polled access — the slightly higher degree of reliability — be a feature end users can enable on their access points? Will it be a feature vendors will use to differentiate themselves? Will we see 802.11a/b/g products “for voice” or “for video”?

“This is the real challenge for vendors and the Wi-Fi Alliance,” Eaton says.

Once 802.11e is completed, you should be able to add it to your existing equipment via a firmware upgrade. Eaton says we’ll first see 802.11e products in the enterprise — in Wi-Fi handsets — then in consumer products, perhaps in the form of consumer grade handsets.

Wi-Fi Alliance

www.wi-fi.org