Generating Reasoning Plan of SWRL Rule with Spark

With the advent of the big data era, large-scale semantic data has emerged. Semantic data contains some important but usually implicit information, which can be derived through reasoning. The conventional single-machine reasoners inevitably have problems such as insufficient computing performance and scalability. And the existing largescale reasoners have limited functions, they cannot fully and effectively support Semantic Web Rule Language (SWRL). In this regard, we propose a scalable distributed reasoning method for SWRL using Spark SQL. This paper shows how to use Spark to generate reasoning plan of SWRL rules, which are used in subsequent reasoning stages. We define a hierarchical data structure to represent the reasoning plan, and designed some variant data structures to facilitate generating reasoning plan. We also use Spark to implement distributed generating the reasoning plan of rule base which composed of rules.