Sleeper: From Cassandra to ScyllaDB to Level the Playing Field
Blog post from ScyllaDB
Sleeper, a fantasy sports app, has positioned itself as a social platform aimed at enhancing user engagement rather than solely focusing on competition. Backed by notable investors like Andreessen Horowitz and sports figures such as Kevin Durant, the app combines advanced statistical modeling with social features to facilitate interactions among friends, family, and colleagues. Originally utilizing Apache Cassandra for its database needs, Sleeper switched to ScyllaDB due to performance issues with Cassandra, particularly its inability to handle materialized views and massive traffic spikes. This transition to ScyllaDB allowed Sleeper to efficiently manage high volumes of data and traffic without the need for a caching system, supporting their rapid growth and expansion into international markets like Europe and South Korea with offerings such as fantasy League of Legends. Using a tech stack that includes Elixir, Kubernetes, and Confluent Kafka, Sleeper effectively manages its infrastructure on Google Cloud Platform, ensuring scalability and reliability.