SRC 3GPP Gateway Peer Communication and Redundancy
The SRC 3GPP gateway configuration allows the definition of multiple Diameter peers to act as PCRFs. The SRC 3GPP gateway picks one peer to become the “active” peer, and forwards all CCR messages to it. If the SRC software detects a failure of the active peer, the SRC 3GPP gateway selects another peer to become the active peer. The SRC 3GPP gateway does not switch back to the original active peer when it comes back online.
For redundancy purposes, you can configure multiple SRC 3GPP gateways to handle the communication between the same group of SAEs and PCRFs. To achieve this redundancy, you configure multiple SRC 3GPP gateway instances to point to the same namespace in the shared configuration.
All SRC 3GPP gateway instances pointing to the same namespace use a Community Manager to elect the active member. The active member registers itself with the naming server and is the only member receiving subscriber-tracking events from the SAE and forwarding them to the PCRF.
All the other SRC 3GPP gateway community members can still receive and process RAR messages from PCRF peers.
If the active SRC 3GPP gateway member fails, the Community Manager detects the failure and elects another member to become the active member. The new active member registers itself with the naming server thus overwriting the old member’s endpoints. During the failure period, the SAE keeps a fail queue for the SRC 3GPP plug-in and replays the queue to the new active member.