Residential and SMB vendors have some 'splainin' to do
I think most who read this blog know that 802.11n is a very big and very complex standard that offers a high degree of variability when it comes to the implementation of products based on its specifications. And I think just about everyone knows that the Wi-Fi Alliance takes a subset of what’s in 802.11n and refines that into a different set of specifications used for interoperability testing and subsequent certification. But even with the amazing technology of the standard and the thoughtful, helpful guidance that approval by the Alliance offers, something (actually, somethings) is still missing – and those somethings are causing confusion among buyers.
The problem primarily centers on the fact that differing product implementations can support between one and four spatial streams, each stream, using a 40 MHz. channel, accounting for up to 150 Mbps of raw throughput. But it doesn’t end there. While the Alliance does test for interoperability, there are many factors beyond simple interoperability that will affect the performance of a given 802.11n product. The other key elements, IMHO, are as follows:
– Is the product single or dual-band?
– Is the product single- or dual-radio?
– Can a given connection automatically shift between bands?
– Can a given radio be disabled (this means you, Apple)
– If dual-radio, is a given radio dedicated to a given band?
– How many spatial streams are supported?
– Is receive antenna diversity supported?
– Is antenna combining (e.g., maximal ratio combining, or MRC) supported?
– Is beamforming supported?
Buyers of enterprise-class products usually know what questions to ask, and suppliers of enterprise-class systems are in general pretty good at pointing out key features and specifications. Not so, however, with many retail residential- and SMB-class products. A recent informal perusal of the stock at my local Best Buy showed that many products provide incomplete or confusing specifications, and some, very surprisingly, don’t even have (or, at least, advertise) Wi-Fi certification.
The Alliance does have an “optional product matrix” which shows the number of streams tested on each supported band. I think it would be a good start to make this information required, rather than optional, and that the Alliance should look into expanding this label with the other information I noted above. Ultimately, of course, it’s a question of throughput, range, reliability, and value (price/performance) – and, of course, Wi-Fi certification is only about interoperability, not performance. What I’m suggesting here, however, is that with the approval of 802.11n, the focus has shifted away from simple interoperability – which one almost always gets whether the finished-goods product has Wi-Fi approval or not – and towards the benefits that advanced architectural and implementation features can provide. I don’t expect the Wi-Fi Alliance to get into the performance testing business, as that group is funded by chip and product suppliers and they obviously for that reason alone can’t be saying which products are best, given, of course, specific test scenarios. But a lot of questions about benefits and not just features are going to come up more and more often now. Consumers really are confused about vendor claims, and the basic Wi-Fi sticker alone isn’t going to be enough to assure customer satisfaction going forward.
