OSQR: A framework for ontology-based semantic query routing in unstructured P2P networks

Abstract

Efficient searching for information is an important goal in unstructured peer-to-peer (P2P) networks. While several P2P systems have been proposed for data sharing purposes, many support only semantics-free keyword searches or coarser grained file name searches. In this paper, we present an ontology based semantic query routing algorithm that performs efficient semantic search in unstructured P2P overlay networks. In our proposed system, the queries are routed in the network by forwarding to peers with highly relevant content in their local storages. To aid in this semantic query routing, we propose a scheme where each peer in the network adheres to a global ontology and semantically tags its local document collection with concepts in the ontology. Based on the semantic tags, peer level semantic summaries are generated, exchanged with neighboring peers and propagated along search paths which aid in efficient local query processing and overlay query routing. An extensive set of simulations performed to evaluate the effectiveness of the system on P2P networks show 380% and 717% improvement in average recall rate, and 410% and 725% improvement in average precision over Ontology Index based Query Routing [20] and Random Walk [18], respectively for dynamic networks at comparable message overheads. Thus, our approach represents a significant advance in practical terms.