With the latest in the bi-annual Wi-Fi Planet conferences (held last week in San Jose) now in the history books, I think it's safe to say that the Wi-Fi space is alive, well, and perhaps never in better shape. The conference served to point out the amazing breadth that Wi-Fi has achieved, and also highlighted upcoming technological and marketing challenges faced by what is still an industry in adolescence.
Oddly, one of the hottest topics was WiMAX, arguably a Wi-Fi competitor and occasionally positioned by some as the eventual replacement for Wi-Fi. There's still significant confusion over the WiMAX that will be based on the fixed-microwave 802.16-2004 standard, and the still-a-work-in-progress mobile version based on the activities of the 802.16e Task Group. The general consensus was that we'll see mobile WiMAX products likely on the order of 18 to 24 months from now.
While most felt that fixed WiMAX would see significant application in public-access Wi-Fi backhaul, especially when deployed in metro-scale mesh architectures, the role of the mobile WiMAX generated significant debate. Steve Saltzman of Intel Capital was quite bullish during the closing analyst roundtable, predicting an aggressive schedule for mobile WiMAX deployment via a strategy of including WiMAX chips in notebook computers something Intel clearly really likes. And one might argue that creating demand-side pull via ubiquitous client capabilities is a great strategy (hey, it's worked for Wi-Fi!).
But mobile WiMAX will require the support of carriers and operators, access to (expensive) licensed spectrum, and face competition from 3G cellular, 4G cellular (not clearly defined yet, but essentially high-speed wireless IP service with VoIP), and Wi-Fi itself when deployed across metro areas " an increasingly frequent phenomenon. The future of mobile WiMAX is thus unclear, but one can't discount the impact of the effort behind it.
Speaking of public access Wi-Fi, meshes were the other really hot topic at the show. Sessions on mesh architectures and technologies were packed, and the core issue for debate was whether multiple radios per mesh node will ultimately be required. With a single radio, of course, operation is store-and-forward a given infrastructure node (mesh access point) can be communicating with a client or another infrastructure node, but clearly not both at the same time.
Francis DaCosta of Mesh Dynamics presented compelling evidence that the accumulated latency in such an approach would quickly result in unacceptable performance. But other panelists in "The Great Mesh Debate" were unconvinced, believing that the relatively light loads typical of public-access Wi-Fi meshes to date could easily be handled with single-radio architectures, and that adding additional mesh nodes would similarly assist here. The issue, of course, is ultimately cost more radios, more expense. But Strix Systems took advantage of the conference to announce its multi-radio outdoor mesh unit, with support for up to six radios. With declining Wi-Fi hardware costs, it does seem that multi-radio meshes will eventually dominate.
Antennas were also a major topic, both on and off the show floor. Multiple-input, multiple-output (MIMO) systems, still new to most in the Wi-Fi community, were addressed both as raw technology as well as in the context of 802.11n. The 802.11n panel reached a consensus (still pretty much a guess, however) that 802.11n-compliant products will hit the market during the second quarter of 2006, ahead of most analyst expectations. Current evidence presented shows a rough doubling of spectral efficiency in today's "pre-n" (a term that still raises hackles among some) products now in the stores.
Beyond MIMO, I was personally intrigued by InterDigital's adaptive interference management (AIM) antenna. This is a tiny, inexpensive electrically-steerable Wi-Fi antenna with associated software designed for interference mitigation. InterDigital is claiming a 100% improvement in (effective) throughput and an 80% improvement in range, and announced Atmel as its first customer for this product.
Motia was also showing its Javelin adaptive-beamforming antenna. Many, including some design engineers, still discount the impact of the antenna on overall radio performance. I like to remind them that the antenna is like the tires on a car the only part of the complex machine that actually touches the medium. As you driving enthusiasts already know, improving the tires can make the car feel like an entirely different vehicle altogether. The same is true in radio.
New Technology and Trends
Returning to our opening theme of the broadening of the Wi-Fi industry, a number of other technologies, applications, and trends were more than evident:
Location and tracking Wi-Fi based location and tracking systems were quite visible, with Ekahau, Newbury Networks, and PanGo Networks hosting good traffic on the show floor. Resolution of these systems can approach a few meters, and even better in some cases. There was a lot of discussion of location-based security as well. Privacy remains a concern, however, as does the need for the client device to be powered up for Wi-Fi based security to work.
VoWiFi VoIP over Wi-Fi was nearly universally hailed as a core driver for the WLAN industry going forward. VoFi becomes particularly important once combined cell/Wi-Fi phones are available, but there are still issues of cost and power consumption to be resolved here. There don't, however, seem to be any technical showstoppers, although there were concerns about the viability of the standalone public-access supplier market. The role that the cellular carriers will ultimately play provided a basis for some debate, especially given the vast amount of free access available in most urban areas. And there was debate over the role of government is enabling (if not operating) public-access Wi-Fi services.
Overall, I left Wi-Fi Planet with the impression that the Wi-Fi space is vibrant, that the enterprise market is growing, and that performance will improve, in terms of throughput, time-bounded services, and perhaps even power consumption. The only complaint was the amount of traffic on the show floor (and maybe the number of exhibitors present), not really all that bad especially considering the relatively limited marketing of the show both nationally and in San Jose.
Craig Mathias is principal of Farpoint Group (Ashland, Mass.)