A massive RDF storage approach based on graph database

Abstract

The efficient management of massive Resource Description Framework (RDF) data is very important in the future of Semantic Web. The RDF data has a natural structure of graph. Thus, the traditional management models such as relational data model and object data model are difficult to meet the requirements of both low data redundancy and high query performance. If it uses a graph model to manage the RDF data, then there is no need to converse the logical data model to the physical data model. This paper proposes a massive RDF storage approach based on graph database. In the perspective of graph model, it partitions the RDF dataset based on a heuristic greedy strategy. It also considers the dynamic property of the data flow. It uses a dynamic mechanism of replicating, storing and deleting the RDF data to achieve load balancing. The experiment is conducted by data storage and query on three datasets. The experimental results show that the proposed approach outperforms traditional methods based on relational database. It also proves that the proposed approach could effectively support the distributed storage of the RDF data.