NetApp’s ‘Project Astra’ Brings Data Management to Kubernetes
Data management and storage provider NetApp wants to help organizations rely on a single platform to manage data services, storage and applications controlled by Kubernetes.
Working with the Kubernetes community, NetApp has introduced Project Astra, a platform to help organizations manage — through single control planes of their choice — the entire data lifecycle for Kubernetes deployments. These capabilities will include disaster recovery, governance and data protection and replication, in addition to full storage capabilities.
A work in progress, NetApp says it is developing, with the Kubernetes community, a platform that will:
- Discover applications for different Kubernetes control planes, whether on-premises or in single or multicloud environments.
- Integrate and unify applications and data management on Kubernetes environments, both in multicloud and on-premises environments.
- Extend the portability for both state and stateless workloads.
- Merge NetApp’s data and Kubernetes management, both as a service and as built-in capabilities.
- Eventually, Project Astra is expected to offer the DevOps community with a software architecture and plugin tools that are compatible with any Kubernetes control planes on offer, the company said, without disclosing a timeline.
Currently, organizations might rely on NetApp’s storage offerings for on-premises and public cloud data management and associated storage volume management. However, Project Astra is under development to help organizations move beyond the management of storage, for example, as a separate and often silo’ed entity. Instead, Eric Han, vice president of product management for NetApp’s Cloud Data Services, told The New Stack that application data management differs from volume data management, since “you have to know the identity of and understand the application and to be able to restore applications and that leads to being able to migrate an application and that leads to portability.”
Project Astra is intended to allow organizations to use the Kubernetes control plane of their choice and to offer comprehensive application data management on Kubernetes and portability for both state and stateless data on Kubernetes, as mentioned above.
“We want people to say ‘Hey, this is where we want to go,’ as opposed to saying, ‘You know what, Kubernetes is just for stateless and that’s all I need Kubernetes for,’” Han said. “That’s a little too small.”
“Organizations could begin using Project Astra as a platform to manage workloads such as database apps and machine learning workloads on Kubernetes clusters and manage entire lifecycle including backup, DR for such workloads,” said Sriram Subramanian, an analyst for IDC to The New Stack. “As of now, these workloads need to use NetApp storage offerings such as OnTap or Cloud Volume Service as the backend,” Subramanian said.
Among the problems Project Astra could solve for DevOps teams deploying on Kubernetes environments is that it could remove much of the complexity from storage and application data life cycle management for DevOps teams working in Kubernetes environments, Subramanian said. “A DevOps person or a Kubernetes cluster admin has little clue on how to manage storage backends, Subramanian said. “Similarly, a storage admin has no clue on the data needs for cloud native apps. Astra attempts to bridge them by abstracting out storage needs at the workloads level.”
The concept behind Project Astra is applicable to small five- to 10-person startups as it is too large multinational companies that might develop machine learning or data science applications or have a number of other different use case scenarios, Han said. Regardless, organizations typically might want to offer what they create as a service to their users or their end-users.
“And they want us to handle the lifecycle,” Han said. “They want us to handle running on their Kubernetes infrastructure.”
On a technical level, a DevOps team is able to select the Kubernetes control panel or interfaces it wants, which Han said “differs from past approaches.” “We’re not saying in order to get started you have to use this and only this,” Han said. Instead, an organization’s choice of Kubernetes clusters will connect to the Project Astra platform, download a set of operators and start to do discovery before beginning the lifecycle management processes.
“Now, you can’t sell that if you don’t have [Container Storage Interface] and have the plumbing, but it’s got to be more than the plumbing,” Han said. “The easiest way I found to explain it is we have to go from storage to app data lifecycle and that requires an understanding of identity, access, etc.”
NetApp is a sponsor of The New Stack.