Multi-Data Source Fusion Approach in Peer-To-Peer Systems

In this paper, a new approach for data fusion in the context of schema-basedPeer-To-Peer (P2P) systems is proposed. Schema-based systems manage andprovide query capabilities for (semi-)structured information: queries have to beformulated in terms of schema. Schema-based P2P are called Peer DataManagement system (PDMS). A challenging problem in a schema-based peerto-peer (P2P) system is how to locate peers who have data relevant to a givenquery.Our proposal lies in application of the multi-data source fusion approach in thecontext of PDMS. Multi-data source schemas, distributed, shared andmaintained by peers, are the basis of a semantic overlay network. The semanticoverlay network and the power of Multi-data source Fusion Language (MFL)are exploited for efficient query routing towards the relevant peers. We showthe design of the Peer Multi-Data source Management System (PMDMS) andwe focus on the Matchmaker and routing components. We give a performanceevaluation of the semantic query routing with respect to important criteria suchas precision, recall, response time and number of messages. We give aperformance evaluation of the semantic reconciliation between peers. Wecompare this result with other systems developed according to peer/super-peerapproach. Finally, we show a prototype developed according to PMDMS. Webuild an application that shares, data between peers, in the domain of leisuresuch as, bank, cinema, restaurant, etc.