Multicast partitioning protocol designs for WDM optical star networks

Abstract

We first propose a simple and efficient multicast partitioning protocol, called Lookback protocol, jor employment in an asynchronous WDM single optical star network. Network nodes use random access schemes to send reseivation packets over a dedicated packet reservation channel. A central traffic scheduler is allocated to coordinate message transmissions. It executes the Lookback scheduling scheme in real time by searching (back into the input queue) for a candidate multicast message that can be received by all of the intended recipients. Such an approach efectively reduces the headof-line blocking effects. Performance results conducted via sinzulations have shown that networks eniploying such scheduling mechanism can exhibit szperior network throughput levels and delay pecformances. To appi'y this inechanisin to a wider range of networking environments, we jirrther consider an interconnected dual-star architecture and propose the Modijied Lookback (Mlookback) mechanism. By using a simple heuristic approach, the Mlookback scheme is able to successjirlly exploit the inter data channels and properly utilize the wavelength reuse proper@ of the intra data channels. Performance results have demonstrated the merits of deploying the underlying Mlookback multicast algorithm in such a dual-star structure.