A Simple Generalized Approach to Node Failure Recovery with Span-Protecting p-Cycles

Abstract

This paper shows that, viewed in a generalized “two-hop” framework for node failure recovery, p-cycles actually have a very high inherent ability to restore paths transiting through a failed node. We also showed that with relatively little, if any, extra spare capacity, the principle is also amenable to explicit design of networks for 100% node and span failure protection with a single efficient set of p-cycles that support both functions. This is very different than the often-prevailing assumption that “ordinary” p-cycles offer no node protection, or only the same node-protection as a BLSR ring embodies. Indeed, the two-hop paradigm for recovery of affected paths transiting failed nodes could provide an attractive option for future network operators in that “ordinary” p-cycles are more localized, fast acting, and simple to plan and operate than any other option such as NEPCs, flow-protecting p-cycles or FIPPs.