Redis Labs Leverages Kubernetes to Automate Database Recovery
Redis Labs today announced it has enhanced the Operator software for deploying its database on Kubernetes clusters to include an automatic cluster recovery that enables customers to manage a stateful service as if it were stateless.
Announced at Redis Day, the latest version of Kubernetes Operator for Redis Enterprise makes it possible to spin up a new instance of a Redis database in minutes.
Howard Ting, chief marketing officer for Redis Labs, says as Kubernetes has continued to gain traction, it became apparent that IT organizations need tools to provision Redis Enterprise for Kubernetes clusters. That requirement led Redis Labs to embrace Operator software for Kubernetes developed by CoreOS, which has since been acquired by Red Hat. IT teams can either opt to recover databases manually using Kubernetes Operator or configure the tool to recover databases automatically anytime a database goes offline. In either case, he says, all datasets are loaded and balanced across the cluster without any need for manual workflows.
Ting notes the biggest challenge organizations face when deploying stateful applications in Kubernetes environments is achieving and maintaining service levels. Given the fact that the Redis database runs in-memory, many IT organizations initially employ Redis to create a persistent cache that enables them to maintain a consistent level of service, he says. Before too long, however, many organizations replace whatever legacy back-end databases they are employing with Redis, he adds.
In fact, as databases evolve, Ting says, they increasingly are being selected and managed by DevOps teams rather than database administrators (DBAs), whose role in the enterprise is starting to diminish.
The core Redis key/value store database runs on a single node. It’s not clear how many organizations that have adopted the open source version of Redis will embrace the enterprise-class edition of the database. However, as the amount of data being stored in Redis databases increases both in the cloud and in on-premises environments, the need for tools to manage instances of Redis across multiple nodes increases. Couple that with the shift to microservices and more complicated infrastructure platforms such as Kubernetes, and the need for Redis Enterprise becomes more apparent.
It’s clear the days when organizations standardized on a single database platform appear to be over. DevOps teams are more comfortable than ever mixing and matching databases as needed. The challenge organizations will face is aligning the type of database employed with the class of microservices-based applications being developed. Like it or not, however, most of those decisions will be driven by developers who are finding it easier to download an open source database to initially build their applications. The challenge that creates for IT organizations is trying to then manage those databases at scale in production environments without having to refactor applications to run on another database. Achieving that goal usually involves embracing a version of that open source database for a fee that is truly enterprise class.
