BFD over Bundle feature enables BFD sessions to monitor the status of individual bundle member links. BFD notifies the bundle manager immediately when one of the member links goes down, and reduces the bandwidth used by the bundle.

There are two types of BFD sessions:

• Single Path Sessions

• Multipath Sessions

BFD can be applied over virtual interfaces such as GRE tunnel interfaces, PWHE interfaces, or between interfaces that are multihops away as described in the IPv4 Multihop BFD section. These types of BFD sessions are referred to BFD Multipath sessions.

As long as one path to the destination is active, these events may or may not cause the BFD Multipath session to fail as it depends on the interval negotiated versus the convergence time taken to update forwarding plane:

• Failure of a path

• Online insertion or removal (OIR) of a line card which hosts one or more paths

• Removal of a link (by configuration) which constitutes a path

• Shutdown of a link which constitutes a path

You must configure bfd multipath include location location_id command to enable at least one line card for the underlying mechanism that can be used to send and receive packets for the multipath sessions.

If a BFD multipath session is hosted on a line card that is being removed from the bfd multipath include configuration, online removed, or brought to maintenance mode, then BFD attempts to migrate all BFD Multipath sessions hosted on that line card to another one. In that case, static routes are removed from RIB and then the BFD session is established again and included to RIB.

In case of BFD multipath sessions, the input and output interface may change based on the routing table updates. If the multipath session BFD packets must get preferential treatment, then a QoS policy must be configured on the entire path, including the possible input and output interfaces of the router.

Note	The CLI bfd multipath include location location is a mandatory configuration to download BFD sessions on a given location.

Bidirectional Forwarding Detection (BFD) is a mechanism used by routing protocols to quickly realize and communicate the reachability failures to their neighbors. When BFD detects a reachability status change of a client, its neighbors are notified immediately. Sometimes it might be critical to minimize changes in routing tables so as not to impact convergence, in case of a micro failure. An unstable link that flaps excessively can cause other devices in the network to consume substantial processing resources, and that can cause routing protocols to lose synchronization with the state of the flapping link.

The BFD Dampening feature introduces a configurable exponential delay mechanism. This mechanism is designed to suppress the excessive effect of remote node reachability events flapping with BFD. The BFD Dampening feature allows the network operator to automatically dampen a given BFD session to prevent excessive notification to BFD clients, thus preventing unnecessary instability in the network. Dampening the notification to a BFD client suppresses BFD notification until the time the session under monitoring stops flapping and becomes stable.

Configuring the BFD Dampening feature, especially on a high-speed interface with routing clients, improves convergence time and stability throughout the network. BFD dampening can be applied to all types of BFD sessions, including IPv4/single-hop, Multiprotocol Label Switching-Transport Profile (MPLS-TP), and Pseudo Wire (PW) Virtual Circuit Connection Verification (VCCV).

/* Verify if the BFD session is up, and the timers are configured. */
RP/0//CPU0:router# show bfd session
Thu Jan 4 03:07:15.529 UTC
Interface Dest Addr Local det time(int*mult) State Echo Async H/W NPU
-------------- ---------- ------------------------ -------- ---- ----- ---
----
Te0/0/0/16.1 172.16.0.1 0s(0s*0) 200ms(100ms*2) UP Yes