On Index Merging for Semantic Web Data

Document Type


Presentation Date

Summer 2010

Conference Name

Proceedings of the Fifth International Conference on Digital Information Management

Conference Location

Thunder Bay, ON

Publication Date


Inclusive pages



In this paper we explore the properties of description indices that store concept descriptions rather than plain data. Although these novel data structures are beneficial for efficiently answering semantic web queries, expressed in a language such as nRQL or SPARQL-DL, they take extra storage and their maintenance can become a performance bottleneck. In order to alleviate these shortcomings, we introduce a procedure for merging description indices. Experimental results over the LUBM benchmark show that this technique can result in economy of storage space, while the performance is slightly affected for a static workload and is improved for a dynamic workload.


Computer Sciences


Link to Original Published Item