$1+{\rm N}$ Network Protection for Mesh Networks: Network Coding-Based Protection Using p-Cycles

Abstract

p-Cycles have been proposed for preprovisioned 1 + N protection in optical mesh networks. Although the protection circuits are preconfigured, the detection of failures and the rerouting of traffic can be a time consuming operation. Another survivable mode of operation is the 1 + 1 protection mode, in which a signal is transmitted to the destination on two link disjoint circuits, hence recovery from failures is expeditious. However, this requires a large number of protection circuits. In this paper, we introduce a new concept in protection: 1 + N protection, in which a p-Cycle, similar to FIPP p-cycles, can be used to protect a number of bidirectional connections, which are mutually link disjoint, and also link disjoint from all links of the p-Cycle. However, data units from different circuits are combined using network coding, which can be implemented in a number of technologies, such as Next Generation SONET (NGS), MPLS/GMPLS, or IP-over-WDM. The maximum outage time under this protection scheme can be limited to no more than the p-Cycle propagation delay. It is also shown how to implement a hybrid 1 + N and 1 + N protection scheme, in which on-cycle links are protected using 1 + N protection, while straddling links, or paths, are protected using 1 + N protection. Extensions of this technique to protect multipoint connections are also introduced. A performance study based on optimal formulations of the 1 + 1, 1 + N and the hybrid scheme is introduced. Although 1 + N speed of recovery is comparable to that of 1 + N protection, numerical results for small networks indicate that 1 + N is about 30% more efficient than 1 + 1 protection, in terms of the amount of protection resources, especially as the network graph density increases.