Case Study: Strava achieves higher throughput and lower latency with ScyllaDB

How Strava’s NoSQL Move Keeps Athletes Moving

Strava, the world's largest sports community with over 100 million athletes, relied on Apache Cassandra for several of its critical data stores. This led to performance challenges, including significant latency spikes and performance issues caused by Garbage Collection (GC) pauses, which impacted the experience for their massive user base. To maintain high availability and improve performance for use cases like activity tracking, Strava sought a more robust NoSQL solution from vendor ScyllaDB.

The vendor ScyllaDB provided a compatible but far more performant database, allowing Strava to migrate use cases like the Horton scalar store with minimal friction. The solution delivered consistently low latencies, eliminated GC pauses, and increased overall throughput. As a result, Strava achieved higher availability for its platform and was able to maintain performance with fewer nodes, which reduced both administrative costs and operational complexity.

Strava

Phani Teja Nallamothu

Senior Cloud Engineer