Case Study: University of Washington achieves fast, Google-like facilities document search with Elastic

University of Washington - Customer Case Study

University of Washington Information Technology (UW-IT) needed to replace a 10-year-old Facilities Services document management system that stores building records dating back to 1861. The existing Oracle WebCenter Content search was rigid and slow—searchable fields and facets had to be predefined and applied across all tenants, reindexing was slow, and users (many untrained) needed a simple, Google‑like search experience with fast, relevant results and tenant-specific access controls.

The team proved out Elasticsearch by indexing ~170,000 documents and building a small microservices stack: a Content‑API that pushes updates to RabbitMQ, an Indexer that feeds Elasticsearch, and a Search‑API that enforces UW access controls. A three‑node ES cluster with minimal mapping tweaks (raw fields, lowercase sorting, search‑as‑you‑type) delivered query times around 50 ms (previously 1–2 s), handled aggregations under 100 ms, and was implemented in about one man‑month. Users reported faster, more accurate search and the solution is now being extended to other customers.

University of Washington

Maxime Deravet

Java software engineer