Case Study: Big Fish achieves real-time personalized recommendations and sub-millisecond latency with MySQL Cluster (MySQL)

Big Fish Selects MySQL Cluster to Serve Real-Time Web Recommendations

Big Fish, the world's largest producer of casual games, needed a way to deliver highly personalized, real-time recommendations to millions of players. The Marketing Management Service (MMS) required very high-velocity data ingest, sub‑millisecond read latency, on‑line scalability, 99.999% availability and enterprise‑level 24x7 support so personalized content could be rendered per page view without impacting user experience.

Big Fish selected MySQL Cluster to leverage existing MySQL skills and meet those requirements: its multi‑master, auto‑sharding architecture, in‑memory tables and SQL/NoSQL APIs enabled high write throughput and low-latency reads. In testing the cluster handled 187,000 inserts/sec while serving 3,000 SELECTs/sec, scaled linearly up to hundreds of millions of records, delivered sub‑millisecond responses, supports on‑line node addition and self‑healing, and now serves the platform’s initial 15M+ sessions/day with 24x7 support and automated cluster management.

Big Fish

Sean Chighizola

Senior Director of Database Administration