Case Study: Orange achieves fast, scalable website search with Elastic

How Elasticsearch Helped Orange to Build Out Their Website Search

Orange’s French search team faced a costly, complex legacy: many vertical search engines built on different proprietary tools and older Sphinx versions, each requiring separate hardware and redundancy. This fragmented stack made it hard to add features quickly while meeting strict requirements for high availability, throughput and low latency.

They consolidated on Elasticsearch (first cluster live end of 2013) for its consistent REST/JSON API and horizontal scalability. Today Orange runs three clusters (largest: 50M documents across 20 VMs, ~150 GB primary index), serving hundreds of requests per second with latencies under 200 ms, simplified client integration and faster rollout of new verticals. Monitoring with Marvel and analytics using Kibana/Logstash support near-zero downtime operations, and most vertical search results on orange.fr are now powered by Elasticsearch.

Orange

Jean-Pierre Paris

Tech Lead