Portworx gets container-native array management for Pure hardware

Pure Storage has announced enhancements to its Portworx Enterprise container storage product that allow it to discover, load balance and manage Pure Storage FlashArray and FlashBlade hardware natively from deployments within Kubernetes.

That is the key upgrade in version 2.8 of Portworx Enterprise, announced at the company’s virtual Accelerate event this week. The addition arguably makes Portworx even more of a container-native persistent storage product than before.

Portworx runs entirely from containers in Kubernetes (and other container orchestrators) and can discover storage and provision and manage it on-the-fly to provide persistent capacity for enterprise applications.

Portworx – acquired by Pure in a $370m deal last year – is effectively a software-defined storage product that runs from containers.

It can build pools of storage, manage its provisioning and provide advanced storage functionality – including backup, disaster recovery, security, auto-scaling and migration – on storage local to Kubernetes cluster servers, on external storage arrays and capacity in the main cloud providers, AWS, Azure and GCP.

Having said that, apart from pools built on local media and now on Pure Storage hardware, storage capacity on external arrays stills need to be configured to be available to Portworx via CSI (container storage interface) drivers.

The capability to configure and manage third-party storage is not likely to be extended to other suppliers’ products, or at least not in a hurry, said Pure Storage EMEA CTO Alex McMullan.

“Historically, storage is not great at multi-vendor cooperation,” he said. “[Extending this functionality to other suppliers] is fine with us and there’s no tech reason why we couldn’t do it, but we will focus first on our own clients.”

CSI is a plugin for Kubernetes and other container orchestrators that allows storage suppliers to expose their products to containerised applications as persistent storage.

At the time of writing, there are about 100 CSIs available for a wide range of file, block and object storage in hardware and cloud formats.

CSI provides an interface between container workloads and external storage that supports the use of persistent storage external to the orchestrator, and advanced functionality such as snapshots and cloning, although it lacks some key enterprise features such as synchronous replication, said McMullan.

The Pure Storage exec was quick to declare CSI as a good thing, but also pointed to the limitations when storage arrays have to manage containerised workloads.

“We have noticed an increase in the scale of and volume of storage requests,” said McMullan. “And at rates greater than an array can handle. We have customers that see 200 volume requests per hour. Storage arrays are not set up to CRUD storage volumes at that rate or volume.”

He added: “CSI is a great building block and we will continue to support it. And CSI won’t become redundant. It has been a good thing, but there are issues around the rate of change.”

Asked about future challenges for container storage, McMullan spoke about the observability of potentially very large and sprawling Kubernetes clusters and the potential for bare metal container deployments to supersede those on VMware.

“As an industry, we need to look at observability, especially in large-scale clusters,” he said. “Kubernetes is difficult to manage.

“Also, containers on bare metal. Over time, and as the Kubernetes security model improves, larger enterprises will be happier with deploying on bare metal, but for now, VMware provides a lot that they need.”

Pure also announced that its predictive analytics platform, Pure1, will now also include container cluster and volume metrics that allow for troubleshooting and resource management.

Leave a Reply