Dan Koepke Lead Platform Engineer at Kwarter
Kirk Stauser Platform Engineer at Kwarter
Kwarter builds a platform and SDKs for real-time data collection and analysis, primarily related to live television broadcasts. We’re part of the platform team, in charge of building the APIs and backend capable of scaling to television-size audiences.
Real-time Messaging with Apache Cassandra
We need high availability, persistence, and scalability for a high-write, low-read workload. Cassandra was the clear choice for our needs.
We use Cassandra for accepting large volumes of messages, such as responses to real-time polls and predictions. We also use it as the store for our social graphs. We are currently on the 1.2.x family running on AWS and are deployed across multiple availability zones.
Cassandra powers an API that sends messages to users in real-time (synchronized with a television broadcast) and allows them to respond.
Getting started with Cassandra and CQL
Make sure you understand the data modeling and data access patterns. And embrace CQL.
We’ve been pretty impressed with how Cassandra has performed under our load requirements and feel confident using it. CQL has made it easier to model solutions for some problems that were tricky using the Thrift interface.