Distributed Execution of Spatial SQL Queries

The volume of available spatial data that is generated and collected has significantly increased in the last few years. A number of applications based on Map-Reduce-like systems and cloud infrastructure have emerged. These applications offer a variety of features, however they differ in terms of spatial functions, partitioning and indexing. In this paper we present our own implementation that enables spatial support for distributed execution of spatial SQL queries as part of the system Exareme. Then, we evaluate some of the State-of-the-Art existing geospatial distributed systems, emphasizing on systems based on Apache Spark. We conduct detailed functional and performance benchmarks that include corner cases that stress the systems in comparison and reveal their advantages and weaknesses in both functionality and performance.