Best practices for processing RDF data using MapReduce (via Hadoop)? - Semantic Overflow