HPE GreenLake For Block Storage Scales Performance, Capacity, Adds 100-Percent Availability Guarantee

‘It fulfills our vision to some extent of delivering a purpose-built cloud-native storage for all workloads. We are talking about workloads from traditional to modern cloud-native, from edge to cloud, allowing us to basically now start looking at how we gain share in this fairly competitive block storage industry,’ says said Sanjay Jagad, vice president of HPE’s product management for block storage.

Hewlett Packard Enterprise Monday unveiled the latest version of its HPE GreenLake for Block Storage built on HPE Alletra Storage MP, which the company called the storage industry’s first disaggregated scale-out block storage, with increased performance and capacity scalability and a 100-percent data availability guarantee for all types of workloads.

The move follows the introduction last April of the company’s move to add new services and hybrid data protection, including a disaggregated scale-out block storage architecture, using HPE GreenLake for Block Storage built on HPE Alletra Storage MP, said Sanjay Jagad, the company’s vice president of product management for block storage.

Last April saw HPE release the industry's first disaggregated block storage solution, breaking with the kinds of storage designs which really have changed little for the last couple decades, Jagad told CRN.

[Related: Storage Vendors Evolve With The Times: 2023 Storage 100]

“I've been doing storage for 25 years,” he said. “The storage designs have not changed. It’s HA (high availability) controllers in a box. And then people making boxes, adding storage at the edge, and cluster them together. The market and the applications and workloads are evolving.”

Businesses are looking for a storage platform that delivers significant value and is very dynamic because those businesses often don’t know what the next workload will be, Jagad said.

Disaggregated storage, Jagad said, is an architecture in which the compute, network, and capacity parts of storage can be independently scaled to meet application requirements.

“For example, if an application requires more performance, can I just add compute nodes only to my scale-out storage without having the need to add more storage JBODs (just a bunch of disks), or vice versa,” he said. “If my application requires more storage, can I just keep adding more capacity? So depending on what the application needs are, I can scale the solution in different dimensions, whether it's compute, storage, or the NVMe fabric to build larger clusters that service providers or large enterprises are looking to build a private cloud on.”

With HPE GreenLake for Block Storage built on HPE Alletra Storage MP release 3, HPE is adding new multi-node switched models that deliver double the performance and 2.5-times more capacity over previous releases, Jagad said.

On the performance side, this includes the introduction of new 2-controller node and 4-controller node switched models offering a choice of 16-core or 32-core configurations, he said. This increases performance by up to double that of the previous 2-controller node switchless models, which limited its use more towards midsized business workloads, he said.

On the capacity side, HPE now supports up to eight JBOF (just a bunch of flash) expansion shelves, each of which supports eight to 24 SSDs per enclosure, that allow non-disruptive scalability from 15.36 TBs to about 2.8 PBs, Jagad said. The SSDs can be upgraded in two-drive increments, and in JBOFs in increments of 1 shelf.

HPE GreenLake for Block Storage built on HPE Alletra Storage MP release 3 also adds expanded AI-based performance reporting and analytics to help businesses simplify their cloud management, a new guarantee that data can be compacted by a factor of four to reduce capacity requirements, and new NVMe-oF/TCP support to take advantage of NVMe over Ethernet networks, he said.

HPE GreenLake for Block Storage built on HPE Alletra Storage MP also allows HPE to offer a 100-percent data availability guarantee as well as predictive performance even if multiple nodes in a cluster are lost, Jagad said.

“We can literally give an experience where this is an always-on infrastructure,” he said. “And what I mean by that is, refreshes across generations of technologies are seamless. You basically now can get into a newer platform. … I can have generations of technologies be introduced, mixed and matched, slowly phasing out the old ones, because there is no state involved in this.”

HPE GreenLake for Block Storage built on HPE Alletra Storage MP is an important milestone in HPE’s journey to change how storage systems are designed, Jagad said.

“It fulfills our vision to some extent of delivering a purpose-built cloud-native storage for all workloads,” he said. “We are talking about workloads from traditional to modern cloud-native, from edge to cloud, allowing us to basically now start looking at how we gain share in this fairly competitive block storage industry.”

HPE GreenLake for Block Storage built on HPE Alletra Storage MP ties into the HPE GreenLake cloud experience that takes advantage of AIOps to help simplify a lot of operational challenges that customers face, Jagad said.

“We have made significant advancements in the way we can simplify this,” he said. “We have an extension of our InfoSight where you now have the ability to do certain AI-related analytics like workload drifts so when workloads come in and you see that there are certain workloads that drift because of certain performance and certain actions that were taken by some rogue users, you see how that impacts and what are the mitigation things that you can do. So new work around AIOps and how it simplifies operations through the GreenLake cloud platform is something you will probably see a lot more.”

Channel partners can use the technology to build a platform that can start as small as a 2U device for mid-market customers and scale for enterprises needing multiple nodes in a disaggregated fashion and help them revamp their data center, Jagad said.

“The same architecture can be used for the edge and at the core,” he said. “Partners now have a way to connect those in an edge-to-core cloud in a true sense. It’s something that can be easily achieved with this architecture.”

HPE GreenLake for Block Storage built on HPE Alletra Storage MP is available to a broad spectrum of channel partners working with small and midmarket or enterprise customers, Jagad said.

“We have a lot of our partners that used to love and sell Nimble Storage [which HP acquired in 2017] who need something now that is midmarket, and we can cater to that,” he said. “We can also work with partners who do stocking and velocity business at the midmarket and entry market. And then we have big enterprise partners or even distributors that are service providers or system integrators that can take advantage of this to scale out and build private clouds or who are in a consulting situation with big enterprise accounts that are planning not only their core data center but also their disaster recovery sites and much more.”

Jagad said the technology can start as low as sub-$35,000 for a 2U unit.