Toward Executing Datalog on Big Data Platforms

Modern databases must be capable of querying massive volumes of data, as we live in the era of Big Data. At the same time, Datalog, a logic language at the roots of database theory, remains a powerful formalism for expressing complex queries in a declarative manner. In this paper, we move the first steps toward bridging these two worlds in an innovative way. In particular, we introduce an approach that compiles stratified Datalog programs in Spark jobs for execution on Big Data frameworks. Our approach aims at effectively combining the declarative nature of logic rules with the scalability of modern distributed systems.