A Network Recovery Strategy Scheme Based on Network Failure Scenarios and Topologies

Network failures happen frequently. There is a need for recovery mechanisms to reduce service interruption. Recovery mechanismsâ advantages and disadvantages are described extensively based on their characteristics and performances. However, it is more desirable that network recovery strategies are chosen based on failure scenarios and topologies. In this paper, we propose a recovery scheme and focus on networks whose paths and resources from source to destination nodes are computed and negotiated primarily at source nodes, Ingress Label Switched Router (LSR), which are the case for Generalized Multiprotocol Label Switching (GMPLS) networks. Choosing proper network recovery mechanisms depends on many parameters such as distance of failure from source node, degree distribution of nodes, availabilities of alternative paths, and maximum allowed-hopcount increase in alternative paths. Three recovery mechanisms: Haskin, Global and Local Protection are compared with the proposed restoration scheme. By changing parameters on appropriate ranges and by using probability of received data packet at the destination node, e.g. probability of error as one of the performance criteria, we can make a fair judgment on choosing a network strategy by considering available network parameters and topology