Summary
This technical report presents the results of performance benchmarks conducted on Stardog version 9.0.1, utilizing various public datasets and benchmarks such as BSBM, LUBM, LSQB, YAGO, and Wikidata. The benchmarks assess transactional queries, reasoning queries, path queries, and bulk loading, while also measuring the impact of concurrent users on system performance, including within a High Availability (HA) cluster setup. Key findings indicate that Stardog can load large datasets at speeds of 1 million triples per second and outperforms a commercial RDF Graph Database in various scenarios, particularly in bulk loading and query execution times. The report details specific performance metrics, including latency and throughput, and compares Stardog's performance against the RDF Graph Database across different concurrency levels. Additionally, it discusses the implications of reasoning at query time versus data load time, highlighting Stardog's efficiency in handling large-scale data operations. The report concludes with detailed results for each benchmark and configuration specifics.
