Configuring BGP4+ Tracking

Easy to deploy, BGP4+ tracking can speed up network convergence and adjust the interval between a peer's being discovered unreachable and the connection's being torn down.

Context

BGP can be configured to detect peer relationship status changes in order to implement rapid BGP convergence. BFD, however, needs to be configured on the entire network, and has poor extensibility. If BFD cannot be deployed on a device to detect BGP peer relationship status, BGP tracking can be enabled on the device to quickly detect link or peer unreachability, implementing rapid network convergence.

Procedure

Access the system view.
system-view
Access the BGP view.
bgp { as-number-plain | as-number-dot }
Enable BGP4+ peer tracking for the specified peer.
peer { group-name | ipv6-address } tracking [ delay delay-time ]

By default, BGP4+ peer tracking is disabled.

A proper value of delay-time can ensure network stability when a peer is detected unreachable.
- If delay-time is set to 0, BGP immediately tears down the connection between the local device and its peer after the peer is detected unreachable.
- If IGP route flapping occurs and delay-time for an IBGP peer is set to 0, the peer relationship between the local device and the peer alternates between Up and Down. Therefore, delay-time for an IBGP peer should be set to a value greater than the actual IGP route convergence time.
- When BGP neighbors successfully perform the GR negotiation, the active/standby switchover occurs on the BGP neighbors, to prevent the failure of GR, delay-time should be set to a value greater than GR period. If delay-time is set to be smaller than the GR period, the connection between the local device and the BGP peer will be torn down, which leads to the failure of GR.