Graph relationship discovery using Pregel computing model

Graph relationship discovery using Pregel computing model Distributed computing is widely used nowdays. Its computational power and memory resources are vital for computations with large-scale datasets, which cannot be handled with a stand-alone system. Pregel is a novel graph distributed computing model, featuring vertex-centric computing, divided into a set of supersteps. In this paper we propose a new algorithm called Pregel Computing Model Algorithm for Relationship Search (PCMARS). Our algorithm is well suited for both directed and undirected graphs, where edges can be weighted and typed. Typed edges are used in RDF graph model for Big Data notion. PCMARS is able to simulate traversal in the opposite direction of the edge, it does not use additional indexing, nor does it change graph structure. We demonstrate PCMARS in our example scenario with the use of Freebase dataset and several test case graphs.