3.9. INTERACTION OF END-TO-END PROTECTION AND FAST REROUTE

So far this chapter has focused on local protection mechanisms using fast reroute. Fast reroute has attractive properties in terms of the speed of recovery, deterministic switchover delay and the ability to protect selected resources in the network. In contrast, using path protection to provide recovery after failure cannot offer similar guarantees, as discussed in Section 3.4.

It would seem therefore that fast reroute makes end-to-end path protection unnecessary. Why compute secondary paths when fast reroute can provide the desired level of protection? If so, are path protection and local protection mutually exclusive? The answer is 'no'; they are complementary.

Path protection allows the operator exact control over the path of the traffic after the failure. Fast reroute has the ability to limit the loss to a few milliseconds. The two can be combined by configuring LSPs with both secondary paths and local protection. Because local protection forwards the traffic around the point of failure, the main LSP can switch over to the secondary path slowly. Furthermore, there is no need to presignal the secondary path and reserve resources for it ahead of time, because there is time to set it up after the failure has happened. On the other hand, the use of a secondary path allows tight control over the traffic patterns following a failure. The secondary can be computed offline taking into account different failure scenarios.

Get MPLS-Enabled Applications: Emerging Developments and New Technologies now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.