To use the structure of networks for identifying the\r\nimportance of nodes in peer-to-peer networks, a distributed linkbased\r\nranking of nodes is presented. Its aim is to calculate\r\nthe nodes� PageRank by utilising the local-link knowledge of\r\nneighborhood nodes rather than the entire network structure.\r\nThereby, an algorithm to determine the extended PageRank,\r\nwhich is called NodeRank of nodes by distributed random walks\r\nthat supports dynamic P2P networks is presented here. It takes\r\ninto account not only the probabilities of nodes to be visited\r\nby a set of random walkers but also network parameters as\r\nthe available bandwidth. NodeRanks calculated this way are\r\nthen applied for content distribution purposes. The algorithm\r\nis validated by numerical simulations. The results show that the\r\nnodes suited best to place sharable contents in the community\r\non are the ones with high NodeRanks, which also offer highbandwidth\r\nconnectivity.
Loading....