Linked data indexing methods: A survey

Abstract

Documents on the contemporary Web are based especially on HTML formats and, therefore, it is rather difficult to retrieve hidden structured information from them using automated agents. The concept of Linked Data based primarily on RDF data triples seems to successfully solve this drawback. However, we cannot directly adopt the existing solutions from relational databases or XML technologies, because RDF triples are modelled as graph data and not relational or tree data. Despite the research effort in recent years, several questions in the area of Linked Data indexing and querying remain open, not only since the amount of Linked Data globally available significantly increases each year. This paper attempts to introduce advantages and disadvantages of the state-of-the-art solutions and discuss several issues related to our ongoing research effort - the proposal of an efficient querying framework over Linked Data. In particular, our goal is to focus on large amounts of distributed and highly dynamic data. © 2011 Springer-Verlag.