Case Study: Hulu achieves uninterrupted, linear-scale streaming to 6M+ subscribers with DataStax

Hulu Relies on Cassandra to Stream Video to More Than 6 Million Subscribers without Interruption

Hulu, an online video service with more than 6 million paid subscribers and availability on hundreds of millions of connected devices, faced a scaling challenge as device usage and real-time user interactions grew. Their legacy infrastructure struggled with the high velocity of writes required to track watch history and resume playback across devices, and a small operations team needed a solution that would scale quickly, ensure continuous uptime, and be easy to maintain.

Hulu evaluated Apache Cassandra, HBase and Riak and chose Cassandra for its linear scale, masterless multi–data-center replication, range-query support and operational simplicity. The result is a primary 32-node Cassandra cluster across two data centers storing several billion rows (≈1 TB unreplicated per data center), enabling seamless cross-device viewing, more than doubled subscribers in two years, and minimal maintenance overhead (one engineer spends <5% of weekly time), with plans to expand Cassandra to additional services.

Hulu

Andres Rangel

Senior Software Development Lead