A client had been advised to replace a number of access points because wireless performance was being blamed on ageing hardware. The proposal sounded plausible: users were complaining, roaming was inconsistent, and the estate had been in place long enough for replacement to feel like a reasonable answer.
The problem was that the evidence did not point at hardware failure. The symptoms were inconsistent by location, time of day and client type. Some users had a poor experience while others were fine. Replacing access points might have improved coverage on paper, but it would not prove why the existing design was failing.
A full wireless survey changed the conversation. The issue was not that the radios were incapable. The issue was channel overlap, excessive transmit power and a design that had become noisy as the environment changed around it. New hardware would have inherited the same behaviour if the design had not been corrected.
The fix was not another purchase. It was an RF redesign: channels, power levels, placement assumptions and roaming behaviour were adjusted against the measured reality of the building.
The complaints stopped because the environment had finally been understood. The vendor recommendation was not malicious. It was incomplete. The hardware was treated as the problem before the radio environment had been proven guilty.
Engineering lessons
- A refresh should not be used as a substitute for diagnosis.
- Wireless problems often live in design, not in the access point model.
- A site survey can prevent a purchase from becoming an expensive guess.