Liveness of SR Policy using Seamless BFD

There are several benefits of performing liveness checks, such as enhanced fault detection, in this context. Once a failing path in a network is detected, immediate corrective measures can be implemented by the operators to avoid any traffic disruption.
This year, we dedicated our testing efforts to evaluating Seamless Bidirectional Forwarding Detection (S‐BFD), which is used to monitor the health of an SR‐TE policy. The given topology comprised two PE routers, each attached to a pair of dual spines. Sessions were created for S‐BFD from an initiator node to a reflector node and sent along the primary path with continuous probes.
To artificially induce failure without physically breaking the links, we applied an ACL on one of the spine links that dropped any MPLS packets (including S‑BFD) but not IGP packets—thus allowing IGP to continue attracting traffic to the failed link. Because S‑BFD is encapsulated in MPLS (the same as the protected SR-TE), it gets dropped, too. The link itself was not taken down; thus, we were able to see how fast the network could detect partial failure by the S‐BFD probes that were not arriving.

Once the S‐BFD detected the absence of returning probes from any instance on the primary path, it triggered the SR‐TE policy to seamlessly switch traffic over to the backup path within the configured failure threshold and delay. Therefore, network operations were not interrupted, which was helpful in the effective functioning of S‐BFD, as it enabled preventive measures in mitigating path failures in an SR‐MPLS environment.