Investigation of a cut-tree approach to network restoration from node loss

The loss of a node in a telecommunications transport network is considerably more severe than loss of a span. The network space capacity will generally be less than that required for 100% restoration of the affected transit demands. We show how a cut-tree analysis of the network can be used to guide network management to ensure that no demands remain completely disconnected and that individual recovery levels are nominally prorated to their pre-failure levels. The new method has been tested with favorable results against "ad hoc" node recovery using a basic path restoration protocol, and against the theoretical benchmark of multicommodity maximum flow. The procedure given has applications in network planning, network management, survivability analysis, and mesh restoration system design.