Today’s enterprises are dealing with vast amounts of data located in disparate storage systems. As enterprises adopt artificial intelligence (AI), data silos exacerbate the challenge of AI training, tuning and inferencing, rendering it a complex and expensive task.

Hewlett Packard Enterprise (HPE) is taking a swing at the pain points of large AI and data lake workloads with new high-density flash options for storing file data.

HPE GreenLake for File Storage is particularly suited for those large AI and data lake workloads because it allows enterprises to consolidate data into one data lake that can be accessed and managed via HPE’s cloud console GreenLake. “This high-capacity, small form factor accelerates the speed at which enterprises can leverage their data for AI initiatives,” HPE Storage VP of Unstructured Data Gokul Sathiacama told SDxCentral.

The file storage’s disaggregated, shared-everything, modular architecture is designed to deliver enterprise performance at AI scale, according to the vendor.

Regardless of where customers are on their AI path, they need quick access to large volumes of data and simple management of those large scale data sets, Sathiacama said. GreenLake for File Storage leverages a disaggregated architecture that helps enterprises scale both performance and capacity of their storage system independently of one another.

With the legacy file storage systems offered by HPE competitors, enterprises need to purchase additional storage systems once they hit a certain data limit. HPE’s file storage, however, lets IT teams scale out storage capacity or performance by adding storage or compute nodes – not by buying another system. “Our competitors cannot match the technological innovation and simplicity that we bring forward,” Sathiacama said.

When compared to the previous release of HPE GreenLake for File Storage, the new options offer four-times the capacity and two-times the system performance per rack unit (RU).

Those improvements translate into the ability to scale throughput at AI scale and reduce the file system’s power consumption by up to 50% compared to the previous release. “Now we can store the same amount of data previously stored in two rack units in a single rack unit, cutting energy consumption in half,” he said.

As AI adoption rises and data centers continue popping up to meet demand, the environmental impacts of these technological advances continues to warrant concern. “Data volumes continue to explode, and a major challenge is how do we store more data while consuming less energy,” Sathiacama said.

HPE claims GreenLake for File Storage lowers data center energy consumption in three major ways: high-density storage, as-a-service consumption and data reduction.

HPE’s high-density hardware, for example, helps “customers to reduce their physical storage footprint and utilize an energy efficient solution in their data center,” he said. “With the new high-density all-flash storage, we have significantly raised the bar again.”

The vendor’s storage-as-a-service model eliminates overprovisioning, or when customers purchase and power more storage than they need. “They only pay for what they consume and add more only when they need it.”

HPE’s data reduction tools use compression, deduplication and similarity reduction to remove unnecessary or redundant data to free up drive space for more efficient use.