A Destination-Oriented Multicast Trees Optimization Algorithm for Controlling P2P Traffic
Yuhui Zhao, Yuyan An · 2009
It is necessary to control the P2P traffic among ASs (autonomous systems) while the P2P applications occupy the most bandwidth of the backbone network. Instead of only restricting the P2P applications, the paper suggests an intelligent relay method based on P2P communities model. The proxy of content caching (PCC) nodes is proposed which are used to manage the local content cache in the P2P communities. The problems of building the core overlay multicast tree (COMT) among the PCC nodes can be reduced to the minimal steiner tree problem (MSTP). For optimizing the multicast trees, the off-line genetic algorithm (GA) is adopted, which uses the destination-oriented method for represent chromosomes and genetic operators. Simulation experimental results show that our model can obtain a near optimal solution for COMT, which has the feathers of low bandwidth cost and high scalability.