RADIUS Redundancy

DIRECT (default): The DIRECT algorithm is designed to primarily focus on redundancy and fault tolerance. In this mode, the RBFS subscriber daemon always attempts to send RADIUS requests to a single preferred server. If a response is received, the server continues to be used for subsequent requests. However, if the server becomes unresponsive or times out, the daemon fails over to the next server in the configured list. This process continues sequentially until a server responds or the list of servers is exhausted.

ROUND-ROBIN: The ROUND-ROBIN algorithm implements load-sharing across all configured RADIUS servers. Each outgoing RADIUS request is sent to the next server in the list, circularly rotating through the servers. This allows the authentication and accounting load to be spread evenly across all available servers.