In order to test the impact to channel utilization on 2.4 GHz of internal APs and Rogue APs, I set up a lab with three APs and went to a series of scenarios, from changing channels and adding bluetooth interference. The expected results are that not only external APs and non-WiFi interference contribute to high channel utilization, but how our internal network was designed as well, even with no clients using the network.
Three APs inside the same room few feet away from each other.
Channel 1, 6 and 11
Power Level 1.
No SSIDs active and so no clients associated.
This is the WiFi environment and Rogue APs. CH 11 is detected at -46 and -57 dBm. CH 6 is detected at -71 dB then CH1 with less power.
No interference from my own APs since they are using 1, 6 and 11. No non-wifi interference detected, other than a FFT pattern. Most activity happens in CH 11 and 12.
Tx (All Power Level 1) and CH assignments (1,6,11).
Channel utilization changes from 0 to 23, being CH 11 the most affected.
Based on this we can determine that even with no SSIDs and no clients connected, rogue APs do affect the channel utilization.
More SSIDs were added, no impact was detected until the fifth SSID was added.
A total of 9 SSID were added.
No clients associated.
Channel 1,6 and 11.
I noticed that channel utilization went up to 7% in some cases.
We can conclude that having more SSIDs does increase the channel utilization but with less impact than the Rogue APs.
All three APs in channel 11, so we can cause more co-channel interference.
No clients associated
Channel utilization was really impacted since all APs are in the same channel seeing each
other at -40 – 50s dBm.
We can determine that our own APs with co-channel interference can impact as much as the Rogue APs.
All three APs in channel 11, so we can cause co-channel interference.
No clients associated
Channel utilization was impacted again in a lower percentage, up to 8% in some cases.
We can conclude that Bluetooth devices do affect the channel utilization, most of the times in all channels, since they are FH devices. The total impact can’t be really determined since they are constantly moving in a real environment, not always active, and the more you have the more inference we get.
From a protocol analysis perspective, we are expecting high control packets traffic, even with or without clients associated. Control packets can consume most of the channel utilization when no clients are connected. So when you see too many mgmt and control packets you can expect a slow network and high channel utilization.
Based on this lab results, we can conclude the following:
Co-channel interference from our own APs or rogues APs impacts the channel utilization by a very high percentage.
Frequency hopping devices or Bluetooth devices impact channel utilization in about 0-8 %. That also depends on how close they are to the transmitting device, and how many they are.
The number of SSIDs does impact channel utilization. Few SSIDs (Less than 4) do not impact too much, but having more than that can impact about 0-7%.
So, to get better performance on 2.4 GHz, we can lower the power levels on our APs in order to reduce co-channel interference, and also try to do the same for rogue APs.
Reducing the SSID number can help, but will not reduce that much the channel utilization.
Removing FH devices will help a lot, but we all know that might not be possible.
In my next post I will show also how to reduce channel utilization by disabling lower data rates.