Efficient Content Location Using Semantic Small World in Peer-to-Peer Networks

Abstract

        Locating content in unstructured peer-to-peer networks is a challenging problem. This paper presents a novel semantic small world resource search mechanism to address the problem. By using vector space model to compute the semantic relevance and applying small world properties such as low average hop distance and high clustering coefficient to construct a cluster overlay. In semantic small world system, the search mechanism is divided into two parts, searching at cluster and outside cluster through inner link and short link, so that it can achieve the incremental research. It significantly reduces the average path length and query cost. Meanwhile, the simulation results show that semantic small world scheme outperforms K-random walks and flooding scheme than higher query hit rate and lower query latency.