Case Study: NASA Jet Propulsion Laboratory achieves real-time, sub-second telemetry search with Elastic

NASA Jet Propulsion Laboratory - Customer Case Study

Jet Propulsion Laboratory (JPL), a NASA field center of ~5,000 staff in Pasadena, faced the challenge of unlocking large, heterogeneous interplanetary datasets—telemetry, designs, docs and collaboration data—while replacing slow legacy workflows (SQL → Excel/Matlab → PowerPoint → email). Missions like MSL and SMAP generate huge volumes (MSL visualizer shows hundreds of millions of values; SMAP produced ~80M telemetry points/day and ~3.8B docs during testing), and JPL needed real‑time, time-accurate search and visualization for engineers and scientists.

JPL built an ETL pipeline (Spark/AWS/lambda), indexed telemetry into Elasticsearch deployed via Mesos/Docker on a hybrid cloud, and added tools like custom admin panels, an adaptive learning graph and the MSL Streams UI. Technical fixes (Top Hits nested aggregation using _source) preserved exact min/max times; the platform supports interactive visualizations and graph queries with sub‑second responses (examples: 24 rows in 240 ms, 92 rows in 147 ms), hundreds of millions to billions of documents indexed, and faster, more accurate analysis and collaboration across missions.

NASA Jet Propulsion Laboratory

Ricky Ma

Chief of Application Architectures