Exploring FHIR Performance Benchmarking
summary
Prototype Python client benchmarks FHIR servers using requests and tinydb, loading 1,000 patients and initialization bundles, running create/read/update/delete five times per resource and summarizing averages and success rates. Multi-threaded runs peak around 130 interactions/sec on test EC2s while single-thread shows best-case response times. Outputs are JSON for per-request timing and job summaries. Tool is not yet open source but the prototype works and highlights database and payload-size limits.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 0:37Why measure performance
- 2:16How to measure performance
- 3:42Prototype requirements
- 5:08Components
- 7:49Workflow
- 9:26Summary Results
- 12:34Improving the Prototype
- 14:07Benefits of FHIR
- 15:12Questions
- 15:36Is your tool open source
- 16:28Have you tested other servers
- 17:48How many max threads
- 18:54Best client in test server
- 20:26Performance metrics
- 23:34payload sizes
- 25:52Outro
