Computing Changesets for RDF Views of Relational Data

Abstract

The Linked Data (LD) initiative fostered the publication of a large number of relational databases in the Linked Open Data (LOD) cloud. A general and flexible way of publishing relational data in RDF format is to create an RDF view of the underlying relational data (RDB-RDF view). An RDB-RDF view can be materialized to improve query performance and data availability. However, an RDB-RDF view must be continuously maintained to reflect dynamic updates over the relational database. The contents of an RDB-RDF view, published on the LOD cloud, can also be fully or partially replicated by other LD applications to improve their efficiency. Thus, updates on an RDB-RDF view must be propagated to maintain other LD application views. This paper proposes a framework for the live synchronization of an RDB-RDF view. In the proposed framework, rules are responsible for computing and publishing the changeset required for the RDB-RDF view to stay synchronized with the relational database. The computed changesets are then used for the incremental maintenance of the RDB-RDF views as well as application views. The paper also describes a strategy for automatically generating, based on a set of mappings, the rules for computing the correct changesets for an RDB-RDF view. Finally, it suggests an architecture and describes an implementation and experiments to validate the proposed approach.