Hulu Scales User Database With Apache Cassandra

As Hulu has expanded the subscriber base to around 5 million, with some 30 million unique viewers per month and some 40 million monthly plays, the over-the-top video provider has been relying on Apache Cassandra distributed database management technology to store and keep track of the viewing history of its users.

Such technologies are particularly important with growing usage of multiple devices, particularly mobile, because they provide real-time access to what a person is watching, explains Andres Rangel, senior software development lead at Hulu.

Using the Apache Cassandra platform, Hulu users can pause one episode on their TV and then resume viewing the same show at the same spot on any other device, such as a smartphone, when they leave the house.

Rangel notes that Hulu selected the Apache Cassandra system two years ago when the previous system was having trouble expanding to its growing subscriber base. “We needed something that could scale quickly and would be easy to maintain because we have a very small team,” he says.

Cassandra beat out a couple of larger competitors and has since been able to easily handle rapidly growing usage. “We have been able to more than double our subscribers in two years and it has had no trouble handling the load,” he explains. “We only need one engineer to maintain it and it takes less than 5% of his weekly time.”

Rangel notes that it has also helped with the database management for adding new services, such features that allow users to control viewing on their Hulu session on smart TVs or gaming consoles with their smart phone. Based on that success, he adds that “we are working to use Cassandra for more services.”