Efficient Linked-List RDF Indexing in Parliament
Dave Kolas, Ian Emmons, Mike Dean · 2009
Abstract. As the number and scale of Semantic Web applications in use increases, so does the need to efficiently store and retrieve RDF data. Current published schemes for RDF data management either fail to embrace the schema flexibility inherent in RDF or make restrictive assumptions about application usage models. This paper describes a storage and indexing scheme based on linked lists and memory-mapped files, and presents theoretical and empirical analysis of its strengths and weaknesses versus other techniques. This scheme is currently used in Parliament (formerly DAML DB), a triple store with rule support that has recently been released as open source. 1