Redundancy

Because the peering relationship between MSDP peers is based on TCP connections, after a failover occurs the TCP connections need to be re-established again.

All SA cache entries learned from the old peering relationships must be flushed and relearned again on new TCP connections.

On a dual MSM system, MSDP runs simultaneously on both MSMs. During failover, the MSDP process on the active MSM receives and processes all control messages. MSDP on the standby MSM is in a down state, and doesn‘t receive, transmit, or process any control messages. If the active MSM fails, the MSDP process loses all state information and the standby MSM becomes active. However, the failover from the active MSM to the standby MSM causes MSDP to loses all state information and dynamic data, so it is not a hitless failover.

On fixed-configuration, stackable switches, an MSDP process failure brings down the switch.