SAS: Semantics Aware Search in P2P Networks

Abstract

Resource management is very important yet challenging in large scale distributed systems like P2P networks. With more and more users incorporating semantic meta-data with their resources, the resource discovery mechanism should not only be able to scale well with the large number of information resources but also be capable of retrieving semantically relevant resources distributed in the P2P network with high accuracy and efficiency. In this paper we address these problems by proposing an ontology-based fully-decentralized peer clustering scheme where the network topology is optimized to perform efficient semantic query routing. The proposed semantic clustering scheme utilizes the structural relationships in ontology to organize peers into clusters based on the semantics of the resources they share. Performance evaluation demonstrates that our proposed approach can dramatically improve the search efficiency of unstructured P2P systems while keeping the communication cost at a comparable level compared with state-of-art unstructured P2P systems.