Wireless Un-Hype and the Urban Wi-Fi Crash

Sorry, but you won't find anything substantive. Instead, the analystvents about interference he experienced in his Boston-based apartment.All in all, the technical and political analyses are iffy at best.

First, the analyst makes the outrageous suggestion that even if youare within six feet of your AP (access point), another AP within 300feet of you can "ruin your surfing or work at home." That's notlikely, unless your wireless gear has some significant bugs. In fact,802.11 wireless networks are remarkably tolerant of the phenomenonknown as co-channel interference. That doesn't mean there aren'tscenarios where the system will die. For example, many reports havesurfaced in recent years about wireless meltdowns at technologytradeshows, where hundreds of vendors try to deploy wireless APs in aconvention center without any effort to coordinate channelassignments. As anyone who has visited a tradeshow will tell you,that's not the real world.

Even in extremely hostile tradeshow environments, 802.11 has performedrelatively well in the face of interference. I queried some engineersat AirMagnet, a leading provider of wireless network analysis systems.These people deal with WLAN issues every day at customer sites. AtAirMagnet's booth at the NetWorld+Interop show in Las Vegas last May,the company detected eight APs operating on the same channel with itsanalyzer. And even though the CRC error rate was 21 percent andchannel utilization was consistently above 30 percent, throughputstill maintained a level of between 800 Kbps and 1 Mbps. Where's thecrash?

Of course there are issues with external wireless interferers,including microwave ovens and other 2.4-GHz devices like cordlessphones, Bluetooth devices and video-monitoring systems. It's alsofairly likely that people in multi-tenant office and apartmentbuildings will experience interference problems if they don'tcooperate with their neighbors. Many vendors acknowledge that WLANtechnology is too complex for most users today, and there's a realneed for chipsets that automatically adjust power levels andtransmission channels to avoid interference. But Aberdeen'spredictions of disaster are largely unfounded, as is the analyst'stechnical analysis of the underlying issues.

The report confuses known, albeit rare, performance problemsassociated with 802.11's MAC design with interference. It goes on tosuggest that a potential solution is for everyone in a building to usethe same 802.11 ESSID. Right. That will solve the problem--not. Theanalyst further warns that doing so will require everyone to "agree ona common security key or password." I have no idea what that means.

The report's other potential solutions include making more RF spectrumavailable, moving to 802.11a and waiting for "smarter technology." Theanalyst asserts that chipmaker Atheros already has extended the rangeof 802.11b to 900 feet at 1 Mbps. Not only is this statement false(Atheros' proprietary new designs extend the range of 802.11asystems), if it were true, it would exacerbate rather than diminishthe problem. Likewise, the proposal to allocate more bandwidth at 2.4GHz doesn't fundamentally solve the problem. If two independentparties set up an AP on channel 1 and the cells overlap, there will beinterference, no matter how many channels are available. Nor does itaccount for the massive installed base of existing radios thatwouldn't be able to take advantage of the additional spectrum anyway.

Aberdeen doesn't stop at offending WLAN technologists. The report alsotakes a shot at FCC policymakers, who have played an admirable roleover the years in the regulation of unlicensed radio. The analystreminds us that these are the same folks who brought us theTelecommunications Act of 1996. Fair enough. Now we all get toremember Aberdeen Group as the organization responsible for this"fine" piece of technology analysis.

Post a comment or question on this story.