On receipt of
the PathErr message, the ingress LSR will switch to the backup path. If the backup
path is presignaled, this can happen within a few milliseconds of the PathErr message
reaching the ingress LSR??”for an overall protection time in the order of hundreds of
milliseconds.
Fast Reroute SONET/SDH networks achieve protection switching in 50 ms for a
ring of up to 1000 km circumference. When running MPLS directly over unprotected
infrastructure (fibres and wavelengths), there may be a requirement to achieve similar
(or better) protection switching at the MPLS layer??”especially when providing circuitbased
services over MPLS. MPLS fast reroute addresses this challenge by repairing
failures at the point of failure, rather than waiting for the PathErr to propagate to the
ingress LSR. When signalling an LSP, the ingress LSR requests fast reroute protection
and indicates whether link or node protection is required. There are two fast reroute
430 Chapter 14
models: one-to-one backup, where each LSR along the path creates a ???detour??? LSP for
each protected LSP, and facility bypass, where each LSR along the path creates a single
???bypass??? LSP for all protected LSPs through the link or node being protected??”generally
to the next hop LSR for link protection or to the next-next hop LSR for node protection.
Pages:
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968