Semantic-Based P2P Model For Heterogeneous Data Management

With the popularity of Peer-to-Peer (P2P) systems, Peer Data Management System (PDMS) has become a promising area in the future interconnection environment. Three key issues related to PDMSs are:(1) Efficiently organizing and managing distributed resources in P2P networks.(2) Accurately and efficiently routing queries from the peer initiating the query to the relevant peers to avoid network flooding.(3) Integrating heterogeneous data returned from different peers to provide users and other peers with a uniform data usage mode.To resolve the above three key issues, this dissertation proposes a semantic-based P2P model. The significant contributions of this dissertation are:(1) The R-Chord model by deploying the Resource Space Model above the Chord overlay for uniformly, normally and effectively organizing and managing resources distributed in P2P networks. Experiments show that the proposed model is more flexible to support semantic-based queries and the average number of nodes involved to answer queries is quite fewer than that of the Chord approach.(2) The Peer-to-Peer Semantic Link Network model by using semantic data links and semantic service links to establish the semantic relationship between peer schemas and between peer services. Experiments show that compared with the flooding and the random walk routing approaches, the recall rate of the proposed approach is close to that of the flooding approach, but the average number of messages generated by the proposed approach to answer queries is the fewest of the three.(3) The semantic-based heterogeneous data integration model, which solves the problem of semantic heterogeneity and structural heterogeneity by schema mapping and query reformulation to integrate the heterogeneous data and provide users and applications with a uniform view.