Systems for big-graphs

摘要: Graphs have become increasingly important to represent highly-interconnected structures and schema-less data including the World Wide Web, social networks, knowledge graphs, genome scientific databases, medical government records. The massive scale of graph easily overwhelms main memory computation resources on commodity servers. In these cases, achieving low latency high throughput requires partitioning processing in parallel across a cluster However, software hardware advances that worked well for developing databases applications are not necessarily effective big-graph problems. Graph poses interesting system challenges: graphs relationships which usually irregular unstructured; therefore, access patterns poor locality. Hence, last few years has seen an unprecedented interest building systems big-graphs by various communities systems, semantic web, machine learning, operations research. this tutorial, we discuss design emerging big-graphs, key features distributed algorithms, as workload balancing techniques. We emphasize current challenges highlight some future research directions.