Ever since Wavion said they'd use space-division multiple access (SDMA) for downlinks in their metro-scale Wi-Fi product, I've wondered whether it could work: A very nice PR flack for the company gave me flak because I compared Wavion's approach to Vivato. In SDMA, you divide client access across space rather than time or frequency. In theory, this allows a single wireless base station the ability to pinpoint a receiver with a unique signal from other, even close by receivers. This reminded me of Vivato's approach, in which a large number of antennas were used to beamform signals; Vivato's product wound up too expensive and not functional enough, and the company went into bankruptcy months ago. (There are new signs of life, however.)

In a recent interview with Wavion for their product launch, their technology head Dr. Mati Wax, explained that they have a distinctly different approach from Vivato, using a single channel and fixed antennas that are digitally controlled. Vivato used three channels and analog beamforming among a large array of antennas. Also, Wavion has production equipment out in the field, while Vivato never got their unique technology to work in production. (If you used a single beam, you could get good results, but the pricing then was out of whack. I still hear about Vivato's panels being used to cover large outdoor areas with low usage rates.)

Wi-Fi isn't designed for SDMA, and thus Wavion can only provide SDMA on the downlink side from base station to receiver. Since routine network traffic like streaming data, Web surfing, and file downloads are all downlink, there's an advantage to a 4:1 asymmetric rate. Further, SDMA would allow specific devices or sets of devices a kind of priority over other devices, making it easier to deliver forms of quality of service (packet priority) in the field.

When I spoke to Wavion's Wax, he said that SDMA support wouldn't be in their first product release, but was still in testing. In an interview earlier this week with Tropos Networks, I asked Saar Gillai, that firm's engineering vice president, whether SDMA was practical. Tropos doesn't use anything like that technology, and Wavion could potentially challenge (or partner with) Tropos if SDMA works as they hope. Gillai said that SDMA was possible--it has been used with specific devices--but fiendishly difficult, which is a fairly positive answer from a company that has gone done a different technology path. Gillai doesn't believe the numbers on density put out by Wavion and others, as he sees MIMO as more of a coverage than performance technology on the metro-scale.

Ultimately, for SDMA to be successful, a future Wi-Fi MAC (media access control) protocol would have to incorporate it so that both adapters and base stations could coordinate more efficiently. The 802.11n spec doesn't include SDMA as an element, so any Wi-Fi update with SDMA that's not a proprietary offering for a niche industry (like military or public safety) would be at least three years away.