Hewlett Packard Enterprise (HPE) released the third version of its GreenLake for Block Storage built on the vendor’s Alletra Storage Multiprotocol (MP) to help customers transition away from legacy node pair data center storage designs to disaggregated storage with multinode switched models.

HPE’s scale-out block storage is managed through the vendor’s GreenLake hybrid cloud platform and offers twice the performance and two-and-a-half times more capacity than the previous generation of this technology, according to HPE Storage SVP Patrick Osborne.

“HPE GreenLake is much more than storage and compute,” Osborne told SDxCentral. It’s a “platform for delivering a hybrid cloud experience across networking, data services, analytics, machine learning and data protection.”

GreenLake for Block Storage furthers HPE’s mission to create a unified customer experience by delivering storage on a common software and hardware stack that can scale to address any size of enterprise.

“We’re able to deliver the same architecture and user experience for data gathering at the edge or massive databases in a data center,” Osborne said. “At a fundamental level, we have radically reduced the complexity and associated cost of scaling data storage and management, making it much easier for us and managed service providers to deliver cloud data services to customers.”

Legacy storage systems are designed with pairs of nodes that can access each other, but not the drives and data connected to a node outside of its pair. That means that if one node fails, the system may lose access to the out-of-commission node’s data and drives.

According to Osborne, these legacy systems are still fairly prevalent today, imposing limitations on data interoperability, performance and capacity across the industry.

To add capacity or performance in a traditional block storage setup, for example, IT teams need to bring in new systems and controllers that would exist in their own storage silo, with unique management and data access rules.

HPE GreenLake for Block Storage, however, allows every node to access every drive and piece of data with an NVMe fabric and does away with the need to buy new hardware to increase capacity.

“When enterprises want to increase performance or data, they no longer have to buy a new system that comes over-provisioned with excess compute or storage,” Osborne said. “Customers can add only the storage JBOFs or compute nodes they require to meet the needs of today’s applications.”

This means data center customers “no longer need to consider hauling in a new system when they want to add a little bit more performance or capacity. Instead, they can add a new compute node or JBOF and do so without interrupting their operations,” he said.

HPE is able to provide a 100% data availability guarantee as a result of its disaggregated, “share everything” architecture. “Even if multiple nodes fail, the other nodes can access connected drives and data through the underlying NVMe fabric,” Osborne said.