Why shall we consider having a LoRaWAN roaming agreement with another LoRaWAN network operator that has partly overlapping network coverage with our network? Our end devices are fixed (e.g.: water meters, temperature sensors, etc.) and our network is slightly larger than our partner's network. Does it make sense to set up roaming in such environment? We operate a ThingPark LoRaWAN network server that can be easily connected to the ThingPark Exchange (TEX) roaming hub.
LoRaWAN has a feature called "macro-diversity", which allows more than one GW to demodulate the same uplink frame (UL).
Multiple gateways receiving the same UL increases network resiliency and performance (in the face of changing/challenging RF conditions).
Thanks to the LoRaWAN Passive Roaming feature, set of multiple GWs demodulating the UL can belong to different networks: home network and one or more roaming partners. (Interestingly, a LoRaWAN device can be seen as at home and roaming into multiple visited networks simultaneously).
This "network collaboration" creates the net effect of multiple networks contributing their GWs to act as "one densified network". Which, thanks to the LoRaWAN ADR algorithm, enables the end-device to use higher data rates with lower transmission power, and therefore reduces the battery consumption and interference. This is a win-win-win for the end-device - home network - visited networks.
That's why it is highly advisable to use LoRaWAN roaming even for the networks with overlapped coverage.