Implementing scalable RDF triple stores that can store many triples and process many queries concurrently is challenging. Several projects have investigated the use of distributed hash tables for this task. This paper gives an overview of the BabelPeers architecture and presents six heuristics to process queries while addressing the issue of network traffic between the peers and data locality in such a highly distributed system. The heuristics are evaluated with the Lehigh University Benchmark.