The vision of the Semantic Web is becoming a reality with billions of RDF triples being distributed over multiple queryable end-points (e.g. Linked Data). Although there has been a body of work on RDF triples persistent storage, it seems that, considering reasoning dependent queries, the problem of providing an efficient, in terms of performance, scalability and data redundancy, partitioning of the data is still open. In regards to recent data partitioning studies, it seems reasonable to think that data partitioning should be guided considering several directions (e.g. ontology, data, application queries). This paper proposes several contributions: describe an overview of what a roadmap for data partitioning for RDF data efficient and persistent storage should contain, present some preliminary results and analysis on the particular case of ontology-guided (property hierarchy) partitioning and finally introduce a set of semantic query rewriting rules to support querying RDF data needing OWL inferences.