The Cloud Native Computing Foundation, which houses the Kubernetes project and other cloud-native technology initiatives, is pitching its latest project Rook as a solution for persistent storage systems in Kubernetes clusters.

Last week the CNCF voted to accept Rook as its 15th hosted project, alongside Kubernetes, Prometheus, OpenTracing, Fluentd, Linkerd, gRPC, CoreDNS, containerd, rkt, CNI, Envoy, Jaeger, Notary and TUF.

Rook is designed to make cloud-native Kubernetes clusters and the applications and services running inside them more self-sufficient and portable, by providing persistent block, file and object storage. Kubernetes is a popular orchestration tool used to manage software containers, which allow applications to be built once and run on any platform.

“Storage is one of the most important components of cloud native computing, yet persistent storage systems typically run outside cloud-native environments,” Chris Aniszczyk, chief operating officer at CNCF, said in a statement.

Building a new system that delivers persistent storage for cloud-native environments would likely take years, CNCF officials said. Instead of creating something new, Rook works by taking traditional storage systems such as Ceph and turning them into cloud-native services that can run on top of Kubernetes.

Persistent storage is desirable for Kubernetes and containers in general, because it means that data can be retained after the applications running inside those containers are shut down. However, most deployments of Kubernetes today rely on cumbersome external storage systems for persistent storage. In public cloud deployments, that typically means using managed services like EBS, S3 and EFS. As for on-premises deployments, these typically rely on storage systems like NFS or traditional information technology storage solutions and hardware.

The lack of native persistent storage has been cited as the biggest challenge of implementing container technologies for two years running in Portworx Inc.’s annual Container Adoption Survey. The survey reported a dearth of adequate tools for managing container storage, adding that worries about data loss are the biggest storage-related issue. Respondents also complain that container storage doesn’t scale very well.

Holger Mueller, principal analyst and vice president of Constellation Research Inc., said that persistent storage was a key requirement for virtually all enterprises, because most of the systems they run require data to be that can be tapped into later for insights about their business and processes. These include systems of engagement, which tell them what customers were looking for, and systems of insight, which provide a glimpse into the consumers themselves, such as where they live.

Companies such as Virtuozzo Inc. have attempted to fill the gap by offering proprietary persistent storage services for container and Kubernetes, but Rook is one of the few viable open-source alternatives that integrates directly with clusters. Essentially what Rook does is provide block, file and object storage services via persistent volumes, while also taking care of tasks such as provisioning and management.

“Not leaving developers to their own integration project with persistence vendors, but offering out-of-the-box cloud native storage, is key to developing next-generation applications fast,” Mueller said.

“It’s a natural fit to run a storage cluster on Kubernetes. It makes perfect sense to bring it into the fold and keep the unified management interface,” said Dan Kerns, senior director at Quantum Corp., the initial sponsor of the Rook project. “With Rook, we wanted to create a software-defined storage cluster that could run really well in modern cloud-native environments, and the storage cluster becomes even more resilient with an orchestrator like Kubernetes.”

Rook is currently available in Alpha, with a beta release and full production-ready version expected later this year.

Image: CNCF