Storing and querying Resource Description Framework.pdf

Optimized Index Structures for Querying RDF from the Web Andreas Harth Stefan Decker Digital Enterprise Research Institute (DERI) National University of Galway, Ireland University Road Galway, Ireland firstname.lastname@ Abstract Storing and querying Resource Description Framework (RDF) data is one of the basic tasks within any Semantic Web application. A number of storage systems provide as- sistance for this task. However, current RDF database sys- tems do not use optimized indexes, which results in a poor performance behavior for querying RDF. In this paper we describe optimized index structures for RDF, show how to process and evaluate queries based on the index structure, describe a lightweight adaptable implementation in Java, and provide a performance comparison with existing RDF databases. 1. Introduction In recent years RDF has emerged as the prevalent data for- mat for the Semantic Web. RDF is a graph-based data for- mat which is schema-less and self-describing, meaning that the labels of the graph within the graph describe the data it- self. Many applications that deal with RDF have the need to store the data persistently and perform queries on the data set. Systems such as Jena2 [17], Sesame [5], rdfDB [7], Red- land [2], Kowari 1 , FORTH RDF Suite [1] and others pro- vide a storage infrastructure for RDF data. However, after looking at the index structures of the systems, most of the systems use an index structure which do not support typical query scenarios for data from the Web which results in poor query answering performance in some cases. Furthermore, to judge a given piece of information we usually need to look at the context of the information. As an example for context consider the source of a piece of information. We might want to trust information coming from more than information from white- (or vice versa). However, the notion of context 1 / is missing in most current RDF storage systems. In this paper we adapt database techniques for RDF