RDF and related Semantic Web technologies have been the recent focus of much research activity. This work has led to new specifications for RDF and OWL. However, efficient implementations of these standards are needed to realize the vision of a world-wide semantic Web. In particular, implementations that scale to large, enterprise-class data sets are required. Jena2 is the second generation of Jena, a leading semantic web programmers' toolkit. This paper describes the persistence subsystem of Jena2 which is intended to support large datasets. This paper describes its features, the changes from Jena1, relevant details of the implementation and performance tuning issues. Query optimization. for RDF is identified as a promising area for future research. Notes: Published in and presented at the First International Workshop on Semantic Web and Databases, 7 September 2003, Berlin, Germany. 18 Pages