Sticky Chromebooks in a 42-AP High School
A 1,200-student high school has 42 APs, all set to maximum transmit power. 5 GHz uses 40 MHz channels; 2.4 GHz uses 20 MHz. 802.11k and 802.11v are disabled, and no minimum RSSI or band steering is configured. Teachers report Chromebooks freezing during Google Meet when classes move between adjacent rooms: the device stays associated to the hallway AP at -80 dBm and 6 Mbps while a classroom AP is at -52 dBm. The controller shows 2.4 GHz channel utilization at 65% and 5 GHz at 28%. A newly installed AP in room 114 is set to channel 36 at 40 MHz and sits 8 m from another AP already on channel 36 at 40 MHz. The principal wants "the newest, fastest settings everywhere."
1. Diagnose the roaming failure and list the configuration changes in the order you would apply them.
Read the worked answer
The client is sticky: it keeps a -80 dBm / 6 Mbps link because the hallway AP's high power makes that marginal signal look usable, and the client has no neighbor information or transition request to move. The 65% utilization on 2.4 GHz and the overlapping 40 MHz channel 36 pair show co-channel interference is also suppressing throughput.
Order matters, because features cannot fix bad RF:
- Fix the channel plan: 20 MHz on 2.4 GHz using 1/6/11; 40 MHz on 5 GHz only where density allows; move AP 114 off channel 36 or drop both to 20 MHz. Never use 80 MHz at this density.
- Reduce power: roughly 8-11 dBm on 2.4 GHz and 14-17 dBm on 5 GHz to shrink cells so a roam happens sooner.
- Enable band steering to move capable clients to 5 GHz.
- Enable 802.11k neighbor reports and 802.11v BSS Transition Management.
- Set a minimum RSSI around -70 to -72 dBm on 5 GHz so clients are nudged off a weak AP; monitor for coverage holes after enabling it.
- Test a real roam while running a ping or video call.
2. Why can lowering AP transmit power improve both roaming and throughput? What must you verify afterward, especially given that client power is unchanged?
Read the worked answer
Lowering AP power shrinks each cell. The client reaches the point where the next AP is clearly stronger earlier in its walk, so it roams at, say, -65 dBm instead of -80 dBm, with a higher MCS and shorter airtime per frame. Smaller cells also reduce co-channel interference, which lowers retries and channel utilization; that is why throughput often rises even though signal strength falls. The principal's request for maximum settings is the opposite of what a dense deployment needs.
Because client transmit power is fixed, the downlink improvement must be checked against the uplink. Verify the RSSI and SNR received at the AP from a representative Chromebook, not just what the client sees; SNR should be 25 dB or better. Also check channel utilization, retry rate, and roaming behavior with a walk test. Watch for asymmetric links where the client hears the AP but the AP cannot hear the client, and for coverage holes created by over-reducing power. If holes appear, add APs rather than raising power back to maximum.