On the Performance of Geospatial Services Deployment: An Experimental Benchmark Analysis

In the rapid evolution of digitization, geoservices such as Geographical Information Systems are re-emerging as a design tool for innovation, change, and policy development, while traditional databases struggle to handle geodata. Scaling these services can be complex, time-consuming, and computationally intensive. Relying on cloud-based solutions, there is a lack of knowledge about (geo-)database performance in the cluster environment Kubernetes, an open-source orchestrator designed to simplify the deployment, scalability, and operation of containerized applications. However, the integration of geoservices into cloud-edge technologies presents challenges in task scheduling and service placement, thus requiring new management strategies. In this respect, distributed computing offers opportunities, but may introduce network delays that degrade the Quality of Service; such latency is minimized by Multi-access Edge Computing standard technology. This research work presents an experimental comparative analysis for PostgreSQL/PostGIS relational database geoservices operating in a cluster versus non-cluster environment, focusing on the data management layer. Performance evaluation refers to the average execution times of SQL operations and the PostGIS function under different hardware conditions. With the combination of Kubernetes and Docker, the clustered environment showed improvements in the average execution time of resource-intensive (geo-)queries on all hardware reference setups, outperforming the non-clustered counterparts.