Optimizing Wi-Fi Performance: Ruckus Wireless’ ChannelFly

As the demands on Wi-Fi networks for throughput, reliability, and capacity continue to mount, upper-level capabilities that optimize radio channel allocation will become even more important than the radio itself.

You might think that, what, with all of the capacity-enhancing innovations in wireless over the past few years, like 802.11n and MIMO, that have improved raw throughput from a megabit or two to 300, 450, and even 600 Mbps, there really wouldn't be much need for cleverness when it comes to boosting performance or capacity further. You'd be, of course, wrong to assume this, and for a number of reasons. First, the behavior of radio spectrum, and thus any given Wi-Fi channel, is statistical in nature - the quality and thus the capacity of any given channel can and will vary, and often from moment to moment, based both on demand (and not just from Wi-Fi sources; interference is almost always a factor today) and the artifacts of radio propagation including most importantly a variety of forms of signal fading. Differences in the transmit power levels among blocks of 5 GHz. channels are another example here. So, then, a little additional intelligence in allocating channels to users and applications can go a long way, and many of the enterprise-class WLAN system vendors have some mechanism for doing just that.

Ruckus Wireless yesterday announced ChannelFly, their latest innovation in this area. ChannelFly, which I have not yet personally tested, is based, according to the press release, on the "statistical analysis of real-time channel performance to discover the true capacity of any channel at any time". OK, there are trade secrets involved here, and Ruckus is of course under no obligation to hand out the blueprints, but a discussion with key Ruckus staff, however, revealed that ChannelFly can make channel (re-)allocations in as little as 15 seconds, and it can achieve allocations that approach 90% of optimal (optimal being defined here as manual allocations based on perfect knowledge of the future). The technology learns over time, and in something between and hour and a day of building its database of channel knowledge, it's doing its job as well as its designers intended.

I have argued for some time that innovations in this area (loosely lumped under the general heading of radio resource management), overall system architecture, and system software and management are much more likely at this point to contribute to overall system performance (as measured by throughput, rate vs. range, and capacity) than improvements in radio technology alone. This assumes, of course that any comparison here is based upon largely equivalent radio implementations, say, two-stream .11n vs. two-stream .11n. And, while we won't see detailed explanations of the algorithms involved in most cases, it will be easy to determine if these are really adding any value, since apples-to-apples comparisons in this case can be done using isolation chambers, and thus won't be subject to the vagaries and variables inherent in freespace testing.

Ruckus builds some great APs; I use one in my office and the performance (rate vs. range) is among the best I've ever tested. The technical team at Ruckus is certainly one of the best in the industry, and ChannelFly is an innovation I'm looking forward to trying myself. In the meantime, it would be wrong to underestimate the potential value of this announcement - as Wi-Fi channels get more congested, techniques like ChannelFly are the best hope for maximizing WLAN performance and reliability.

• Wi-Fi
• capacity
• ChannelFly
• performance
• radio resource management
• reliability
• Ruckus
• Ruckus Wireless
• wi-fi
• Wireless LAN
• wireless LANs
• WLAN
• WLANs