Skip to main content

Comcast on Sprinting from Cassandra to ScyllaDB

Comcast’s X1 platform supports more than 30 million set top boxes and another 15 million more “second screen” devices used on a monthly basis. Their X1 Scheduler processes over 2 billion RESTful calls daily. To meet that scale, the X1 Scheduler uses multiple datastore technologies, including Cassandra, MongoDB, Elasticsearch and ScyllaDB. Learn how ScyllaDB was introduced into the Comcast NoSQL ecosystem earlier this year, and how they anticipate saving over 60% of their Cassandra operating costs once the migration is complete. Learn the methodology Comcast used through their evaluation, risk analysis and testing processes, plus details of their benchmarking, migration and deployment plans.

Comcast’s Philip Zimich, Senior Director of Software Development & Engineering at Comcast,  leads the architecture, development and operations of the Comcast’s X1 Scheduler system. The X1 Scheduler powers the DVR and program reminder experience on the Comcast X1 platform, a cable and streaming video service that supports more than 31 million set top boxes and “second screen” devices used on a monthly basis for 15 million households. Their X1 Scheduler processes more than 2 billion RESTful calls daily. To meet that scale, the X1 Scheduler uses multiple datastore technologies, including Cassandra, MongoDB, Elasticsearch and ScyllaDB.

As Zimich shares in the above video, the Xfinity X1 platform saw such dramatically improved performance after migrating from Cassandra to ScyllaDB, especially on long-tail latency, they were also going to rip out the caching layer for its UI. The performance ScyllaDB provides is “a massive, massive gain compared to where we are today.”

Zimich  also ventures back in history to Comcast’s move from Oracle to NoSQL via Apache Cassandra, its subsequent shift from Cassandra to ScyllaDB. He shows how ScyllaDB helps the Xfinity X1 team save on their costs and reduced their node count by an order of magnitude (from 962 Cassandra nodes to only 78 ScyllaDB nodes), while ensuring their user base still has ample room to grow.