An optimal cost and high efficient traffic grooming algorithm for large WDM networks

Abstract

Wavelength division multiplexing (WDM) is a key technology that makes a massive increase in optical network bandwidth. The emerging problem is how to optimize the optical network capacity efficiently. Traffic grooming is an efficient technique to utilize network resources by packing low speed traffic streams onto a high-speed lightpath, using cross-connections of conversions between different transport systems or layers within the same system [1]. Most existing traffic grooming algorithms have high complexities and do not consider costs paid for the systems. To improve these limitations, this thesis introduces a simple but very effective traffic grooming approach called time and cost efficient traffic grooming algorithm (TCGA). The proposed algorithm spends two stages. First, I reduce the complexity of an algorithm by removing redundant nodes in the process of building reachability graph. Then I propose the cost functions to determine the optimal grooming path for the new connection request. The simulation results show that my algorithm has a lower complexity compared with existing ones, an optimal cost and an acceptable blocking performance.