Case Study: Pokémon Go achieves scalable, low-latency player profiles with Couchbase

Keeping viewers happy with high performance and high availability

Pokémon Company International’s Pokémon Trainer Club— the legacy profile service behind Pokémon Go and other Pokémon apps—faced a sudden scalability crisis when the game’s launch pushed user numbers past Twitter. Built on SQL Server, the relational system couldn’t keep up with the low-latency updates, growing queries, and rapidly evolving player profiles required for a global, real-time gaming experience.

The company migrated to Couchbase Server, using N1QL and Multi‑Dimensional Scaling to scale query and index services independently (avoiding costly rebalances and data moves) and leveraging memory‑optimized indexes and cross‑datacenter replication for high availability. The new profile system was developed in under three months, delivered consistent low-latency performance at scale, and minimized operational cost as usage grew.

Pokémon Go