RDF support over relational and non-relational sources - overview
A number of use cases within IBM have recently driven a requirement for an efficient and scalable persistence mechanism for RDF data. We explored whether it was possible to build an efficient RDF store on top of relational sources (DB2 and PostgreSQL) to exploit some of the existing strengths of the relational database -- such as concurrency, compression etc. (see SIGMOD 2013 'Building an Efficient RDF store over a relational database, and https://github.com/Quetzal-RDF/quetzal), and extend it to non-relational stores such as Apache Spark. We have also extended this approach to property graphs (see supplementary information).