Survivability of optical networks for high availability

In this thesis, solutions are proposed for the survivable routing, line rate and wavelength assignment problem in mixed line rate wavelength division multiplexed optical networks. The focus of the solutions is to reduce the communication cost and the traffic blocking by increasing the effectiveness of resource allocation. The work in this thesis is divided into three parts. The first of these is optimal resource allocation (OPT) method which is a multi objective optimization method using multi level programming, focusing on the resource selection part of the SRRWA. The second one is dynamic line rate assignment (DYN) method which considers the daily fluctuations of aggregated traffic on backbone networks. In the third part of the work, SecondPass (SP) method and its learning based variations SPL and SP+L, which dynamically assign line rates to lightpaths considering individual requirements of each traffic request, are proposed. The aim of the work is to propose solutions which allocates optimal resources for the traffic in consideration by also increasing resource utilization for the benefit of future traffic requests. By considering single link failures, survivability is ensured using shared and dedicated path protection techniques. The performance analysis of the proposed methods is performed on various network topologies under different scenarios.