Design and analysis of survivable telecommunications networks

In this study, we consider the design and analysis of two types of survivable telecommunications networks, namely, mesh restorable networks and self-healing ring (SHR) networks. We compare the two architectures under various demand patterns and various network sizes.;Traditional survivability measures in survivability analysis of networks have focused on network connectivity, however, it is important to incorporate traffic into survivability analysis of fiber optical networks. The propose a class of traffic-based survivability measures, which calculate the average traffic-carrying capability of a network, over all possible failure states, for demand between any pair of nodes, within the network, or within a subnet. We consider a preplanned restorable mesh architecture in the analysis, and in order to evaluate the measures without specifying rerouting for each failure state, we present an upper and a lower bound of the measures and describe a procedure to calculate the bounds in polynomial time. For self-healing ring architectures, we derive closed form expressions of the traffic-based measures.;The major part of this study is on the design of survivable networks. We present a generic cost model, in which the cost of any link consists of the installation cost and the material cost, for the design of both mesh and ring networks. The design of mesh restorable networks consists of three stages, network configuration, working capacity assignment, and spare capacity assignment. We combine the first two stages to minimize the cost before restoration is added into the network. The enumeration method which searches for the optimum configuration is computationally feasible only for small networks, we present a configuration algorithm which starts from the fully connected network and gradually removes links to reduce the cost while maintaining the connectivity requirement. For spare capacity assignment, we consider path restoration since it achieves higher capacity efficiency. Spare capacity is added to a link to protect against any single failure and is reused for different failures. We also address a joint approach which combines the three stages together.;To utilize self-healing ring technology in large networks, we propose single-homing and dual-homing hierarchical self-healing ring (SH-HSHR, DH-HSHR) architectures to interconnect self-healing rings. The HSHR architecture consists of multiple levels of single SHRs and a higher level ring is used to connect several lower level rings. We also consider the configuration and dimensioning of HSHR networks. We show that the enumeration method can only be used for small networks due to the complexity, and present heuristic algorithms to find near-optimal HSHR configurations. Dimensioning of a HSHR is transformed into dimensioning of single rings inside the HSHR. A routing algorithm for bidirectional SHRs is presented to complement the shortest path routing. Finally, we present numerical results and compare the two restoration strategies, mesh and ring in terms of survivability and cost.