NetApp Unveils AI Storage Architecture, Oracle Cloud Service And Agent Controls
NetApp announced a storage architecture for AI data centers, a storage service with Oracle Cloud Infrastructure and updates to its data platform Wednesday. The company framed the products around one problem: AI agents act on enterprise data at machine speed, while the infrastructure underneath was not built for that pace.
Storage And The GPU Problem
The centerpiece is NetApp Novus, an architecture that NetApp says is designed to exceed throughput of 100 terabytes per second and support hundreds of thousands of GPUs. Arindam Banerjee, NetApp's chief platform and technology officer, said storage stopped limiting applications years ago because those applications ran on CPUs. GPUs changed the equation, and new GPU generations arrive every 12 months.
Banerjee said market studies showed GPU utilization as low as 5% at some deployments. The company's press release puts the figure below 30% for storage designs that predate AI factories. Banerjee described the arithmetic behind...
Copyright of this story solely belongs to www.ciol.com. To see the full text click HERE