Performance Comparison of Spatial Data Indexing Using Distributed Systems

In this study, the efficiency and scalability of Geomesa and Mosaic, two well-known platforms for spatial data indexing and join operations in distributed environments, are investigated using Apache Spark and Scala/Python. The experiments investigated the impact of varying the input sizes, index utilization, and multi-node environments. Geomesa exhibits commendable performance, particularly without geohash indexing incorporation, demonstrating its efficiency in handling large-scale spatiotemporal datasets, as Geomesa requires 10.6% less time than Mosaic. By contrast, Mosaic employing the H3 indexing methodology consistently outperformed Geomesa, achieving 16.6% less time. These findings underscore the pivotal role of spatial indexing, with H3 indexing emerging as a superior technique in join operations. This research provides valuable insights into the comparative strengths and efficiencies of Geomesa and Mosaic, providing practitioners with informed perspectives for spatial data processing applications in distributed systems.