Which option is a critical mechanism to optimize convergence speed when using MPLS FRR?

Down detection, typically accelerated using BFD (Bidirectional Forwarding Detection), determines how fast the head-end or PLR (Point of Local Repair) learns of a failure and activates the pre-computed FRR backup path. Without fast failure detection, the pre-computed backup provides no convergence benefit because the failover trigger is delayed.

Shared Risk Link Groups (SRLGs) improve backup path diversity and reliability but do not directly accelerate the speed at which convergence occurs after a failure.

Bandwidth reservation ensures capacity on backup paths but has no bearing on how quickly a failure is detected or the FRR switchover is triggered.

IGP timer tuning affects overall topology reconvergence, but MPLS FRR is specifically designed to operate independently of IGP convergence - the bottleneck is detection speed, not IGP.