During Pure Storage’s Accelerate event in Las Vegas this week, the company announced a significant upgrade to its Fusion control plane. This upgrade will allow storage capacity to be easily provisioned to specific applications, particularly those using artificial intelligence (AI) workloads.
One of the key announcements at the event is the implementation of storage classes that enable seamless provision of storage capacity across Pure Storage arrays. This enhancement is aimed at targeting storage for AI applications or those utilizing AI technology. Additionally, Pure Storage introduced AI Copilot, which offers natural language technical support for Pure arrays.
Another notable announcement is the launch of Evergreen//One for AI, a service that promises storage performance guaranteed based on bandwidth requirements of GPU clusters. This is done in accordance with a service-level agreement (SLA). The company also unveiled Secure Application Workspaces for Containers, which offers secure multi-tenancy and policy-based governance for Kubernetes container management. This integration allows for seamless communication between critical data and AI clusters.
Pure Storage also anticipates receiving Nvidia DGX SuperPod certification by the second half of 2024, having already been certified for Nvidia DGX BasePod and OVX. CEO Charles Giancarlo envisions Pure Storage as evolving into a comprehensive data management storage platform. The company is focusing on enabling self-service through Fusion, allowing IT to create APIs that expose storage classes for easy accessibility.
The Fusion control layer, introduced in 2021, has now been enhanced to provide backward compatibility, enabling access and management of block and file storage across the entire fleet from a single array. This feature set will be available for block and file storage in 2024, with object storage to come in the future.
According to Pure Storage’s vice president and general manager for core products, Shawn Hansen, Fusion’s latest capabilities address the evolving landscape of enterprise storage. The addition of backwards compatibility for existing arrays and datasets has been a significant development. This new functionality will be provided as an update with no additional licensing costs.
Furthermore, Pure Storage’s AI Copilot, which is currently in tech preview with customers, utilizes GPT-4 and a vast training data corpus to provide natural language support. The company is working towards further enhancing customer experience by incorporating additional information for training purposes.
Overall, Pure Storage is focused on providing innovative solutions to meet the evolving storage requirements of AI-driven applications and workloads.