Measuring WLAN Performance: Ruckus Goes Open Source

A step towards a tool that really measures wireless performance

As you probably know, I spend a lot of time benchmarking and otherwise testing wireless LAN systems. This is a tricky business; I've written widely on the topic, and the bottom line here is that it is very challenging in a freespace environment to definitively compare two different systems. We can of course use the same benchmark tools and traffic/load parameters in each case; the process here is essentially identical to what we'd do in the case of wire. But wireless introduces two additional variables over which one doing the testing has no control:

* First of course, in the nature of radio propagation itself. This is a fundamental, physical (as in physics), unpredictable, statistical, non-linear phenomenon that's essential to the universe itself. Good luck dealing with that one beyond long run times, multiple runs, and using turntables where possible to factor out antenna orientation (although, in the case of beamforming, such obviously isn't a good idea). But physics is physics and care is thus required in doing comparative testing

* Second is the nature of the 802.11 protocols themselves. Every product vendor, or at least WLAN chip vendor, has their own unique take on how best to use these, and drivers additionally allow a degree of variability in a number of key parameters. Add in settings like transmit power, and you have the makings of a multi-week exercise that might still not yield definitive results. But we can get close or at least derive an inference or two in most cases.

So, taking the above into account, I have become a fan of VeriWave's WaveAgent tool, and, of course, Iperf (version 1.7.0 in my case), which is no longer being maintained but which is regardless simple, flexible, and open-source. But I'm always interested in other approaches, and some time ago I had a discussion with the folks at Ruckus Wireless about an internal tool they'd developed, called Zap. Zap is particularly oriented around the analysis of video traffic, with an emphasis on both throughput and time-bounded behavior. Very importantly, it was designed with wireless, and particularly wireless LANs, in mind. And, again to their credit, Ruckus has now released Zap as open source, and is to be commended for making an important contribution to the testing and performance-evaluation community.

I have a copy of Zap and will give it a try in the video testing project we're now doing. I am, BTW, still looking for a tool that can align 802.11 behavior with actual throughput at any given moment in time. This could be done by correlating packet-trace information with the output of Zap or another tool. Hey, with Zap now open source, it may very well become the platform for many future innovations along these lines.

OK - now I'm off to CES...

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Now read: Getting grounded in IoT