RF interference has some notable characteristics: amplitude, frequency width, and airtime utilization. It’s pretty normal among Wi-Fi infrastructure systems for an AP to continually measure RF interference and then react, if necessary, by changing channel. Whether the AP reacts or not is based on threshold values set by the manufacturer, and when the AP reacts is often configurable by the administrator.
With this information as a baseline, it’s the “measure RF interference” part that I’d like to address. Most systems today measure RF interference based on amplitude and frequency width, but the more sophisticated systems also look closely at airtime utilization.
When speaking in “spectrum analysis” terms, you might hear the related lingo “Duty Cycle.” When a device is transmitting all of the time (never takes a break), it is said to have 100% duty cycle. While 802.11 devices aren’t allowed to have 100% duty cycle (because they must receive acknowledgements, listen before they speak, and only get to transmit a small amount of data before relinquishing the channel).
It takes a group of very active clients to push channel utilization to very high numbers. It usually takes non-802.11 devices, that continually transmit at high power, to completely kill a channel. Below is an example of an OFDM-based Wi-Fi device configured to transmit random data at high power continually, ignoring 802.11 rules. This shows 100% duty cycle and use of all of a channel’s airtime.
It’s entirely possible to have a hideous-looking (on a spectrum analyzer) RF interference source that isn’t negatively impacting your AP’s performance all that much due to a minimal amount of airtime usage. I think I’m stating the obvious when I say that we don’t want unnecessary AP channel changes due to how it affects applications.
Wi-Fi is surprisingly resilient. Of course, it ought to be with all of the overhead built into the PHY and MAC layers of the protocol! With Bluetooth, Zigbee, cordless phones, security cameras, microwave ovens, and a dozen other common device types living in the 2.4 GHz band, Wi-Fi usually continues to crank away so long as there is sufficient airtime left on the channel.
The worst real-world interference source I’ve witnessed in my tenure as Wi-Fi Geek has been 2.4 GHz security cameras that use 100% of the airtime on 2.4 GHz channels. I’ve seen low-power units and high-power units, but all of them cause channel death due to airtime consumption.
Anybody want to share additional information along these lines?