BGP fast reroute in IP-VRF network instances

Note: This feature is supported on 7250 IXR Gen 3, 7250 IXR Gen 2, and 7250 IXR Gen 2c+ platforms.

BGP fast reroute (FRR), or Edge Prefix Independent Convergence (PIC), combines indirection techniques in the forwarding plane and pre-computation of BGP backup paths in the control plane to support fast reroute of BGP traffic around unreachable or failed BGP next hops.

When BGP FRR is enabled, BGP attempts to program a backup path for every IPv4 or IPv6 non-ECMP prefix in the route table that is derived from imported EVPN or IP-VPN routes.

For more information about BGP fast reroute, see the SR Linux Routing Protocols Guide.

Configuring BGP fast reroute

When backup-paths install true is configured, BGP finds and programs a backup path for every non-ECMP prefix within the scope of the command. A BGP-VPN route is a backup path candidate if it is valid, its BGP next-hop is reachable and its BGP next-hop is different than the BGP next-hop of the primary path. The backup path is the best of the candidate paths and selected using the same BGP decision process and tie-breaks that are applicable to primary path selection.

The following shows the configuration of BGP FRR for IPv4 VPN routes. For configuration that applies to IPv6 VPN routes, use the ipv6-unicast command.
--{ +* candidate shared default }--[  ]--
A:srl1# info network-instance test protocols
    network-instance ip-vrf {
        protocols {
            bgp-vpn {
                backup-paths {
                    ipv4-unicast {
                        install true
                    }
                }
            }
        }
    }