Dynamic Multicast Traffic Grooming in Optical WDM Mesh Networks: Lightpath Versus Light-Tree

Abstract

With the rising popularity of multicast applications, various algorithms using either lightpath or light-tree schemes have been proposed for dynamic multicast traffic grooming in meshed wavelength division multiplexing (WDM) networks. To the best of our knowledge, however, no systematic comparison has ever been made between the performance of the two schemes in minimizing network blocking probability. In this paper, we address the dynamic multicast traffic grooming problem in WDM networks and present comprehensive comparisons between these two schemes in different cases. Our main contributions are twofold: first, we compare the performance of the existing lightpath- and light-tree-based grooming algorithms and show that, in most cases, the lightpath-based methods outperform the light-tree-based ones. We discuss and explain such observations. Second, we propose a lightpath-based algorithm, called the LightPath Fragmentation (LPF) method, to further improve the network blocking performance. Numerous simulations show that the LPF method steadily outperforms the existing algorithms in different cases. Effects of the ratio of unicast traffic loads versus overall traffic loads and the average number of destinations of each multicast request are also studied.