The switch uses one BFD session for all protocols with the same destination. For example, if a network runs OSPFv2 and BGP across the same link with the same peer, only one BFD session is established, and BFD shares session information with both routing protocols.
You can enable BFD over data paths with specified OSPFv2 and OSPFv3 neighbors, BGP neighbors, static routing next-hop addresses, and Fabric Extend tunnels.
The switch supports BFD asynchronous mode, which sends BFD control packets between two systems to activate and maintain BFD neighbor sessions. To reach an agreement with its neighbor about how rapidly failure detection occurs, each system estimates how quickly it can send and receive BFD packets.
A session begins with the periodic, slow transmission of BFD control packets. When bidirectional communication is achieved, the BFD session comes up.
After the session is up, the transmission rate of Control packets can increase to achieve detection time requirements. If Control packets are not received within the calculated detection time, the session is declared down. After a session is down, Control packet transmission returns to the slow rate.
If a session is declared down, it cannot come back up until the remote end signals that it is down (three-way handshake). A session can be kept administratively down by configuring the state of AdminDown.
In asynchronous mode, detection time is equal to the value of DetectMult received from the remote system multiplied by the agreed transmit interval of the remote system (the greater of RequiredMinRxInterval and DesiredMinTxInterval.) DetectMult is approximately equal to the number of sequential packets that must be missed to declare a session down.