WLAN Analysis, Troubleshooting and Optimization
Tool categories
Separate the tools by what they see. A protocol analyzer in monitor mode captures 802.11 frames and exposes retries, management behavior, roaming and the 4-way handshake. A spectrum analyzer sees the physical layer - microwave ovens, cordless phones, Bluetooth, non-Wi-Fi interferers and channel width occupancy. A survey tool maps coverage. Wireshark decodes captures; controller dashboards aggregate client health.
Metrics that matter
Retry rate above roughly 10–20% signals poor SNR, interference or a sticky client at the cell edge. CRC/FCS errors point to collisions or interference; a low data rate with high airtime consumption points to a coverage problem, not a capacity problem. Channel utilization above about 50% is a warning sign; above 70% expect voice quality loss. Throughput is not data rate - expect roughly half of the PHY rate in real TCP traffic because of half-duplex operation, headers, acknowledgments and contention.
Method and traps
Work bottom-up: physical layer (power, antennas, interference), then layer 2 (channel plan, retries, roaming), then layers 3–7 (DHCP, DNS, mDNS, RADIUS timeouts, QoS). Common question traps: blaming DHCP for slow throughput when retries are high; assuming 5 GHz always fixes a problem; confusing band steering with load balancing; ignoring sticky-client behavior and 802.11k/v/r support; misreading a spectrum plot as a protocol capture.
Optimization levers
Right-size channel width, disable low legacy rates to shrink cell size, tune transmit power so cells overlap rather than blanket, enable band steering and 802.11k/v/r, enable WMM for voice, and use multicast-to-unicast conversion for mDNS-heavy environments.
Worked example
A 300 Mbps PHY rate with 40% retries plus 60% channel utilization will yield far less than a clean 150 Mbps link. Fix the RF first: reduce interference or narrow the channel, then re-measure.