Distributed Storage For Massive RDF Data Based Don Graphatabase

Efficient massive RDF data management is a key part of the semantic web technologydevelopment. Due to the RDF data model has graph feature, the traditional management stylebased on relational data model or object-oriented data model are difficult to meet the low dataredundancy and high query performance simultaneously when facing large-scale RDF data. If wemanage the RDF data based on graph model, not only we can avoid the transformation fromRDF logical data model to physical data model, but also we can use mature graph algorithm tooptimize the RDF data query. So this paper proposed a massive RDF data distributed storagesolution based on the graph database.Firstly, this paper introduced some current RDF data management styles as well as theiradvantages and disadvantages, and discussed the applicability of management method based ongraph data model for the current large-scale RDF data. Then paper introduced some basictheoretical knowledge related to our topics. Then we proposed the overall architecture of ourmassive RDF data distributed storage system based on graph database and discussed the designdetails and related algorithms of the main modules in our system. Finally in the experimentalpart, the results showed that the overall performance of storage solution based on graph databaseis superior the storage solution based on relational database. Also the experiment verified thedistributed storage characteristics of our system. This paper has the following innovations:1)Thedistributed extension for graph database.2) Partitioning the RDF dataset from graph data modelto ensure the data locality.3) Implementing the conversion from SPARQL query language tograph query language.