NVM Express Announces the Release of the Computational Storage Feature

Computational Storage Will Standardize Vital Storage Management Features Across the Cloud, Enterprise and IoT Industries

NVM Express, Inc. announced the release of the NVM Express (NVMe) Computational Storage Feature, which provides a standardized, vendor-neutral framework for connecting applications to NVMe Computational Storage devices across both compute and storage services. Computational Storage includes two new command sets: Computational Programs and Subsystem Local Memory Command Sets.

“Computational Storage is a standardized approach that enables an open, interoperable ecosystem. By offloading compute to the device, we anticipate that these industries will experience reduced total cost of ownership and overall performance boosts.”

“NVM Express Computational Storage is part of our efforts to help enterprises and hyperscale data centers meet the ever-evolving demands of the storage industry,” said Bill Martin, NVMe Computational Storage Task Group Co-Chair and Board Member. “Computational Storage is a standardized approach that enables an open, interoperable ecosystem. By offloading compute to the device, we anticipate that these industries will experience reduced total cost of ownership and overall performance boosts.”

CIO INFLUENCE News: North American Manufacturers Go Digital to Navigate Changes

Computational Storage reduces data movement by providing computation on a storage device. Performing computation on the device leads to faster response times for latency-critical applications including databases, AI suggestion models and content delivery networks.

CIO INFLUENCE News: Apiiro and Akamai Launch Technical Alliance to Deliver Code-to-Runtime API Security

Key NVM Express Computational Storage Features:

• The Computational Programs Command Set provides a host-driven, modular approach to computational programs.
• The Subsystem Local Memory Command Set provides the ability to access memory in an NVM subsystem via computational programs and via an NVMe transport, accessible through NVMe I/O commands.