cnPilot access points support wifi meshing to allow easy extension of the coverage area, without worrying about running more cables. This article will cover some of the common best practices while deploying WiFi Mesh.
Avoid single points of failure: whenever possible enable two Access Points as Mesh Base units, for each of the Mesh Clients in your deployment. This way if one of them goes down the Mesh client will have an alternate path back to the network, instead of the one link loss resulting in network connectivity loss for all.
Enable security on the mesh links: When a mesh link is configured it can be setup for open, or wpa2-pre-shared-keys, we would recommend that administrators use wpa2-pre-shared-keys which would provide AES encryption.
Prefer 5GHz links for mesh: except in remote areas the 2.4GHz spectrum is fairly well saturated so the 5GHz spectrum tends to be cleaner than 2.4GHz, and provides more capacity (wider channels, higher MCS rates). An exception to this might be fairly longer distance links where 2.4GHz might provide the range and penetration across walls that 5GHz may not.
Consider ePMP, PMP and PTP for outdoor: while the WiFi mesh is a quick and easy way to get wireless connectivity, in very high interference, performance critical or very long range connections Cambium has several other products that are more suitable for wireless connectivity. With APs such as the E501 providing PoE out to power these devices, installation and cabling can also be simplified even if there are two devices on a pole/tower.
Prefer non-DFS channels: if setting static channels, or when using automatic channel selection set the preference to non-DFS channels, this way mesh links would come up faster (Avoid the time spent on radar scan which is needed on DFS channels), radar detections will avoid taking down the complete mesh briefly as channels are changed, and typically the allowed transmit power tends to be higher on non-DFS channels.
Keep mesh line-of-sight: as far as possible try to keep the mesh base and mesh client in line of sight of each other, for optimal performance.
Consider using the mesh radio for client access too: the cnPilot mesh implementation does not require the radio doing mesh to be dedicated to just mesh, it can do double-duty and serve up WLANs for client access at the same time. Depending on your applications and use case it may be beneficial to keep both client and mesh enabled on the radio. In other cases, the performance impact (since the airtime is now split between client and mesh) may suggest the use of keeping the radio dedicated to mesh.
Avoid more then 2-3 hops: while cnPIlot APs support multiple hops, with each mesh hop there is atleast a 50% drop in throughput (since a radio in the middle of the chain has to first receive then re-transmit the same packet upstream, using the air-time on it twice). A single hop design is typically ideal with exceptions for 2-3 hops.