FlexTable: Using a dynamic relation model to store RDF data

Abstract

Efficient management of RDF data is an important factor in realizing the Semantic Web vision. The existing approaches store RDF data based on triples instead of a relation model. In this paper, we propose a system called FlexTable, where all triples of an instance are coalesced into one tuple and all tuples are stored in relation schemas. The main technical challenge is how to partition all the triples into several tables, i.e. it is needed to design an effective and dynamic schema structure to store RDF triples. To deal with this challenge, we firstly propose a schema evolution method called LBA, which is based on a lattice structure to automatically evolve schemas while new triples are inserted. Secondly, we propose a novel page layout with an interpreted storage format to reduce the physical adjustment cost during schema evolution. Finally we perform comprehensive experiments on two practical RDF data sets to demonstrate that FlexTable is superior to the state-of-the-art approaches. © Springer-Verlag Berlin Heidelberg 2010.