As MPLS is increasingly deployed to support high revenue services such as Layer 2 connectivity, IP VPNs and voice and video services, there is correspondingly an increased pressure on the availability of MPLS networks. This talk focuses on various techniques to improve this availability. Much has already been said on fast reroute to protect against link failures; here, we will primarily discuss methods of protecting against LSR failures, including graceful restart and stateful replication. There will also be a brief description of Bidirectional Failure Detection, a recent development that speeds up detection of link or control plane failures.