How it Works

In the preceding diagram, NRF 1 is Primary and NRF 2 is secondary for SMF. On bringing up, the SMF registers (NF registration) with NRF 1 and starts NF heartbeat with NRF 1. The SMF uses the heartbeat response to track the operational status.

In case, the SMF detects NRF 1 failure by missing NF heartbeat response, the SMF registers to NRF 2 (secondary NRF) and starts sending NF heartbeat. The SMF continues to send NF Register message1 to NRF 1 to keep track of its status.

If the SMF receives register response from NRF 1, it detects that the NRF 1 is up again. The SMF marks NRF 1 as active once it recovers and stops sending NF heartbeats to NRF 2.

Note

NF Reregistration (default behavior) on failover and fallback should be configuration driven. When NRF 2 detects that the SMF has stopped sending heartbeats, it checks from NRF 1 if it has received SMF registration by using discovery with SMF instance ID.

As the management and discovery endpoint groups are separate, the Registration based operation status check is not used for NRF failure handling during NF discovery. During NF discovery, the configured NRF endpoints within the group are attempted in the priority order. If the first choice NRF endpoint is not responding the next best NRF endpoint is chosen and so on.