Graph Processing at Oracle, A Pragmatic and Holistic Approach
In this talk, we present our on-going efforts for building graph processing systems that are well integrated with existing robust data management solutions. We first discuss the benefits of using efficient representations for graph data analysis and review two different kinds of graph workloads as well as existing systems designed for each workload. We will outline current challenges in graph analysis and then compare and contrast two graph runtimes that have been implemented by Oracle Labs: a scale up, single-node, multi-threaded engine and a scale-out distributed runtime. We show that by adopting a set of high level Domain Specific Languages (DSLs), one can migrate between both runtimes seamlessly so that the best engine can be used for a particular dataset size. We conclude by outlining interesting areas for research and further development.
Hassan Chafi is Director, Research & Advanced Development at Oracle Labs where he currently leads various projects. His research investigates high-performance, parallel, in-memory Graph Analytics and using domain specific languages (DSLs) to simplify parallel programming. Dr. Chafi received his PhD from Stanford University. His thesis work at Stanford focused on building a Domain Specific Language Infrastructure, Delite. He was advised by Dr. Kunle Olukotun. Prior to that, Hassan worked in the area of hardware transactional memory as part of the Transactional Coherence and Consistency (TCC) project at Stanford where he developed a scalable extension to the original TCC protocol.