Abstract: A communications packet network comprises a plurality of nodes interconnected by communication links and in which tunnels are defined for the transport of high quality of service traffic. The network has a set of primary traffic paths for carrying traffic and a set of pre-positioned recovery (protection) traffic paths for carrying traffic in the event of a fault affecting one or more the primary paths. The network incorporates a fault recovery mechanism. In the event of a fault, traffic is switched temporarily to a recovery path. The network then determines a new set of primary and recovery paths taking account of the fault. The traffic is then switched to the new primary paths. The new recovery paths provide protection paths in the event of a further fault. The network nodes at the two ends of a recovery path exchange information over that path so that packets returning to the main path present labels that are recognizable for further routing of those packets.

Abstract: A system, device, and method for bypassing network changes in a communication network pre-computes recovery paths to protect various primary paths. A fast detection mechanism is preferably used to detect network changes quickly, and communications are switched over from the primary paths to the recovery paths in order to bypass network changes. Forwarding tables are preferably frozen as new primary paths are computed, and communications are switched over from the recovery paths to the new primary paths in a coordinated manner in order to avoid temporary loops and invalid routes. New recovery paths are computed to protect the new primary paths.