Speed Up Queries in Unstructured Peer-to-Peer Networks

Unstructured peer-to-peer networks have gained a lot of popularity due to their resilience to network dynamics. The core operation in such networks is to efficiently locate resources. However, existing query schemes, e.g., flooding, random walks and interest-based shortcut, suffer various problems in reducing communication overhead, and shortening response time. In this paper, we study the problems in prior works, and propose a new query scheme by mixing inter-cluster queries, and intra-cluster queries. Specifically, the proposed scheme works by efficiently locating the clusters sharing similar interests with inter-cluster queries, and then exhaustively searching the nodes in the found clusters with intra-cluster queries. To facilitate the scheme, we propose a clustering algorithm to cluster nodes that share similar interests, and a labeling algorithm to explicitly capture the clusters´s borders. As demonstrated by extensive simulations, our new query scheme can improve the system performance significantly by delivering a better tradeoff between communication overhead and response time.