With millions of customers to serve, a major cable company needed to build a virtual customer support agent to scale their customer requests. The app had to provide 24/7 help to users across the United States, and store metadata about customer conversations.
The first version of the application was built on Amazon Aurora, in a single cloud region on the east coast of the United States. However, this deployment was vulnerable to failures, and when a networking failure in an AWS region knocked the entire service offline, the team realized that Aurora’s single-master architecture was not sufficient to attain the always-on customer experience they wanted. They needed to explore other options in hopes of migrating the app.
>Here’s what they were looking for in a new database:
- Always-on availability, because a 24/7 virtual agent can’t have downtime, even in the event of a failure.
- Strong consistency, meaning the data should always be up to date, no matter what machine or data center handles a request
- Low latency reads, from anywhere in the US
They turned to CockroachDB for its strong resiliency, consistency, and low-latency reads. The end result is a multi-region deployment that can survive datacenter failures:
Head to the full case study for a graphic of the implementation pattern they’re using today, complete with how they’ve configured their cluster to ensure both local reads and ultimate survivability.
