2 min Applications

Pure Storage, Nvidia and Cisco launch AI hyperscale systems

Pure Storage, Nvidia and Cisco launch AI hyperscale systems

Pure Storage has combined its AIRI artificial intelligence (AI) platform with Nvidia’s DGX-1 and DGX-2 hyperscale systems, and rolled out a new FlashStack for AI with Cisco and Nivida. That’s what ZDNet reports.

Specifically Hyperscale AIRI Nvidia DGX-1 and DGX-2 systems and Pure Storage FlashStack for AI. The Hyperscale system has Infiniband and Ethernet as interconnect options, and the systems can grow across racks.

Pure Storage unveiled AIRI, which stands for AI Ready Infrastructure, last year to simplify the AI workload infrastructure. AIRI is a turnkey infrastructure for large-scale AI applications. This means that cloud, enterprise and government organisations need to gain faster insights into how to deliver innovations. AIRI was realised together with Nvidia.

FlashStack for AI is a converged system developed with Cisco and Nvidia for AI workloads. The system consists of Cisco UCS C480ML and Pure Storage FlashBlade, as well as Nvidia GPUs and AI libraries. Both FlashStack for AI and Hyperscale AIRI Nvidia DGX-1 and DGX-2 systems work with Nvidia’s NGC software container registry and various machine learning frameworks, as well as Kubernetes.

Collaborations with Nvidia

Pure Storage is the latest in a series of companies working with Nvidia to create hyperscale systems for AI and machine learning. For example, last week Dell EMC announced that it had developed a reference design for AI workloads with Nvidia. This architecture combines Dell EMC all-flash Isilon F800 with the Nvidia DGX-1.

A major reason storage providers are collaborating with Nvidia on AI systems is because many AI deployments so far have taken place on premise. “The primary reason people choose on premise is a property problem. A lot of data is already on premise or is generated on premise. Moving data to the cloud is a problem,” said Brian Schwarz, vice president of product management at Pure Storage.

“My first advice is to build your AI system where your dataset is. If it’s all in the cloud, use the cloud. If data comes from a factory or hospital, build the system there.”

This news article was automatically translated from Dutch to give Techzine.eu a head start. All news articles after September 1, 2019 are written in native English and NOT translated. All our background stories are written in native English as well. For more information read our launch article.