cid 1d08bfffd46c62d7905c7556b6bcc4f06aeb708d4f4a80c6ea274019a332f60d · original · open access · via openalex
Title: TimescaleDB with PostGIS versus MongoDB for Agricultural Database Workloads Authors: Mustafa Elkomy; Erick Noboa; Verőné Wojtaszek Malgorzata; György Eigner Venue: IEEE International Symposium on Intelligent Systems and Informatics (SISY) Year: 2026 Canonical: doi:10.1109/sisy71351.2026.11738543 URL: https://doi.org/10.1109/sisy71351.2026.11738543 Collector: openalex OA: true ABSTRACT Precision irrigation decision support systems require the storage and querying of high-frequency weather observations alongside spatially indexed field data and raster-derived vegetation indices. This study benchmark TimescaleDB (PostgreSQL+PostGIS) against MongoDB on an application-driven mixed spatiotemporal workload derived from an operational irrigation DSS, measuring ingestion throughput, storage footprint, and query latency under controlled Docker resource constraints. On 500k-record datasets, MongoDB ingests data about 2.5× faster (12012–13815rec/s vs. 4810–5655rec/s) and uses 35% less disk (168.4MB vs. 259.8MB). TimescaleDB delivers lower median latency for analytic and spatial queries (e.g., 59.7ms vs. 238.6ms for a spatial range aggregate) and wins three of five query types, while MongoDB is faster for point lookups and BLOB retrieval.