Clustering using Hypergraph for P2P Query Routing - Simulation and Evaluation
Résumé
Peer-to-peer overlay networks offer a flexible architecture for decentralized data sharing. In P2P schema-based systems, each peer is a database management system in itself, ex-posing its own schema. In such a case, the main objective is the efficient search across peer databases by processing each incoming query without overly consuming bandwidth. The usability of these systems depends on efficient and effective routing of content-based queries is an emerging problem in P2P networks. This work was attended to motivate the use of mining algorithms in the P2P context to improve the efficiency of such methods. Our proposed method combines clustering and hypergraphs. We use ECCLAT to build approximate clustering and discovering meaningful clusters with slight overlapping. We use the algorithm MTMINER to extract all minimal transversals of a hypergraph (clusters) for query routing. The set of clusters improves the robustness in queries routing mechanism and scalability in P2P Network. Our experimental results prove that our method generates impressive levels of performance and scalability with respect to important criteria such as response time, precision and recall.