Abstract:Span protecting p-cycle has been shown to be a promising approach for survivable WDM network design because of its ability to achieve ring-like recovery speed while maintaining the capacity efficiency of a mesh-restorable network. In [12], the concept of span protecting p-cycle is extended to path-segment protecting p-cycle (flow p-cycle for short) which can protect a multi-span segment of a working path. An ILP model that computes the optimal placement of flow p-cycles for protecting a given set of demands (i.e., static traffic) is given in [12]. In this paper, we present an algorithm that uses flow p-cycles for service protection in dynamic traffic scenario. When a demand arrives at the network, a working path needs to be selected for the demand and a set of flow p-cycles need to be configured for protecting the demand. To utilize the network capacity efficiently, we propose to reuse the existing flow p-cycles in the network to protect the current demand. Based on this idea, we develop an ILP formulation to compute the working path and flow p-cycles for the current demand to minimize the working and spare capacity required by the demand. We also describe a demand teardown procedure that accurately computes the network capacities that can be reclaimed when a demand departs the network. Simulation results show that flow p-cycle outperforms span p-cycle to a considerable extent for dynamic traffic.

Dynamic Provisioning of Survivable Services Using Path-Segment Protecting P-Cycles in WDM Networks