Portworx has made available a software development kit (SDK) for its OpenStorage storage software for containers, giving DevOps teams more control over when, where and how data is provisioned and managed when they deploy an application.
In addition, Portworx is now also adding support for VMware vSphere and Pivotal Container Service (PKS) via version 1.6 of PX-Enterprise, as part of an effort to extend the reach of its storage software into the realm of private clouds.
Portworx CTO Gou Rao says the company developed the SDK in response to requests from organizations that had embraced DevOps for more programmatic control over storage. The SDK provides access to an application programming interface (API) that can be invoked to build workflow tools enabling DevOps teams to provision storage resources and perform tasks such as backup and restore without any intervention from a storage administrator. The OpenStorage SDK is based on an open source software that, according to Rao, has the largest number of contributors to an open source storage project.
In fact, Rao notes, storage administrators who typically rely on tools that have a graphical user interface have become a chokepoint within many DevOps pipelines. The SDK eliminates the need for DevOps teams to ask a storage administrator to perform tasks that developers now can perform whenever required, he says.
The OpenStorage SDK enables DevOps teams to use native calls over gRPC using languages such as Golang, Python and others, or invoke a RESTful API. Tasks such as allocating, consuming, encrypting, backing up and managing storage; defining I/O service levels; creating snapshot; and migrating data can be automated.
Rao notes that capability now can be extended to private clouds typically deployed in an on-premises environment, in addition to existing support for a variety of container services provided by cloud service providers. Most organizations building stateful applications now want to be able to deploy those applications across a hybrid cloud computing environment, he says.
Rao notes that in general, there has been a sharp increase in the number of organizations building stateful applications using containers that require access to databases and persistent storage. Much of that spike in development can be attributed to Kubernetes clusters becoming a de facto standard. In fact, Rao says, the ecosystem surrounding Kubernetes is now maturing at a rate that greatly exceeds previous open source platforms such as OpenStack.
It’s not clear to what degree storage administrators have become a bottleneck as organizations seek to embrace more sophisticated approaches to DevOps. In many cases, storage administrators are performing a gateway function to determine which applications accessing a limited number of storage resources should have more priority over others. But in may scenarios there are instances in which a containerized application is accessing dedicated storage resources that could be managed more easily within the context of a larger set of DevOps processes.
It may take a while before the internal IT politics surrounding storage in the enterprise get resolved. But as SDKs for managing storage become more widely available, it’s now only a matter of time before the issue gets forced one way or another.