Wi-Fi Infrastructure Power Testing - Part II

When I published my last report on Wi-Fi infrastructure power consumption, I received an important criticism - that, since I didn't actually measure the power being consumed, it's possible that all four of the power sources I used were in fact putting out way more than the 15.4 Watts allowed under 802.1af, and that the AP that was supposedly compliant with .3af might not in fact have been such. While the probability of this being the case seemed well less than .5, such criticism was (and remains) indeed valid. And, while I'm not in a position to certify .3af compliance (again, we're not a compliance/conformance lab), I set out to correct this obvious deficiency, even though I was confident that the results would show primarily the same results.

And, so, Extreme Networks volunteered to let me use their new Summit WM200 WLAN controller and Altitude 450 and 451 APs, along with their X450e-24p gig-E switch, which is capable of measuring the actual power being consumed at each switch port. They also let me use their absolutely wonderful Faraday cage, perhaps the best I've ever seen. For those of you who do this kind of testing, however, my advice is to dress lightly. We've still not mastered the concept of air conditioning within such an enclosure, and it gets pretty hot over the hours of elapsed time required to obtain 90 seconds of test data.

But we did in fact sweat through a whole bunch of tests, the results of which you can find in our latest Farpoint Group Tech Note. And the results here are in fact substantially the same as before - properly designed, a .11n AP can indeed run at very high throughput, two radios simultaneously blazing, using nothing more that .3af power.

I really hope this closes the books on this topic. And, since I wrote the first Tech Note, the price of electricity has climbed even higher - a lot higher in some locales. The incentives for and reality of infrastructure-size power conservation are now, I believe, well-established.