YouTube brings Vitess MySQL scaling magic to Kubernetes

YouTube is working to integrate a beefed up version of MySQL with Kubernetes

YouTube is working to integrate a beefed up version of MySQL with Kubernetes

YouTube is working to integrate Vitess, which improves the ability of MySQL databases to scale in containerised environments, with Kubernetes, an open source container deployment and management tool.

Vitess, which is available as an open source project and pitched as a high-concurrency alternative to NoSQL and vanilla MySQL databases, uses a BSON-based protocol which creates very lightweight connections (around 32KB), and its pooling feature uses Go’s concurrency support to map these lightweight connections to a small pool of MySQL connections; Vitess can handle thousands of connections.

It also handles horizontal and vertical sharding, and can dynamically re-write queries that could impede the database performance.

Anthony Yeh, a software engineer at YouTube said the company is currently using the service to handle metadata for the company’s video service, which handles billions of daily video views and 300 hours of new video uploads per minute.

“Your new website is growing exponentially. After a few rounds of high fives, you start scaling to meet this unexpected demand. While you can always add more front-end servers, eventually your database becomes a bottleneck.”

“Vitess is available as an open source project and runs best in a containerized environment. With Kubernetes and Google Container Engine as your container cluster manager, it’s now a lot easier to get started. We’ve created a single deployment configuration for Vitess that works on any platform that Kubernetes supports,” he explained in a blog post on the Google Cloud Platform website. “In this environment, Vitess provides a MySQL storage layer with improved durability, scalability, and manageability.”

Yeh said the company is just getting started with the Kubernetes integration, but once users will be able to deploy Vitess in containers with Kubernetes on any cloud platform supported by it.