Digree: Building A Distributed Graph Processing Engine out of Single-node Graph Database Installations

Abstract

In this work we present Digree, a system prototype that enables distributed execution of graph pattern matching queries in a cloud of interconnected graph databases. We explain how a graph query can be decomposed into independent sub-patterns that are processed in parallel by the distributed independent graph database systems and how the results are finally synthesized at a master node. We experimentally compare a prototype of our system against a popular big data engine and show that Digree provides significantly faster query execution.