In this thesis, we propose a novel and efficient heuristic Routing and Wavelength Assignment (RWA) algorithm, called Wavelength Scarcity Avoidance and Lambda-Run-based (WSALR) algorithm for wavelength-routed WDM networks with limited-wavelength-converting capacity. Lightpaths for all connection requests in the network can be set up and released on-demand. The performance of the proposed algorithm is intensively studied by analysis and computer simulation. Three important performance dimensions considered here are the blocking probability of a call, the setup delay, and the fairness.
The collected statistical results show that the proposed algorithm outperforms other previously proposed algorithms. Its lower blocking probability enables networks to sustain much higher traffic load, typically about 118% to 260%, and its lightpath setup phase can be accomplished within a shorter time. Besides, the simulation also denotes that the WSALR exhibits a better fairness, i.e. the difference between blocking probability of the shortest routes and that of the longest routes is mitigated. These advantages justify that our proposed algorithm can be feasible for large dynamically-reconfigurable next generation optical networks.