Hello!
I did the deployment last weekend with 20 fat AP 650 (single radio!) In an event for 1000 people, where 700 of them used the Wi-Fi.
With low density, my retry index was spinning around 7% (acceptable). During peak hours of people, I saw through the NSight an increase of 40-50%. From what I've seen, even the device with Wi-Fi not connected, it causes pollution of beacons (probes circulating in the network).
I did the whole project on Ekahau, using the concept of small cells, minimum bit rate of 18 Mbps, 802.11 k / r and the best resource of all: load-balance group of clients between APs. It worked! haha !. That's where WiNG's big differential comes in!
Unfortunately I was not in place during the event, I was not able to do Layer 1 & 2 troubleshooting (Chanalyzer, OmniPeek, etc). I was curious to debug the frames, make adjustments in the beacons (DTIM interval), and follow the evolution in NSight ...
Anyway, from the comments I received from the users, I had no serious complaints.
Thank you!


Contrib:
Wi-Fi / WLAN Performance Management and Optimization