Businesses across nearly every industry are adopting artificial intelligence (AI) to stay competitive, drive growth, and decrease expenses. From self-driving vehicles to genomic medicine, chatbots, and predictive maintenance, the success of AI depends completely on access to large amounts of data that can be used to identify patterns, develop predictive insights, and enable increasingly accurate autonomous systems.
Almost 12 months ago, NetApp and NVIDIA partnered to create NetApp® ONTAP® AI to offer customers a turnkey solution for supporting AI, machine learning (ML), and deep learning (DL) workloads with enterprise-class performance, reliability, and support. NetApp ONTAP AI integrates DGX servers, NVIDIA V100 Tensor Core GPUs, and NetApp cloud-connected all-flash storage systems with state-of-the-art networking.
ONTAP AI unifies data management across the data pipeline with a single platform that enables users to securely control and protect their critical data across edge, core, and cloud. The solution is fully integrated and is purpose-built for the performance and scalability that AI environments demand. It supports seamless, cost-effective movement of data across premises and clouds. ONTAP data management software delivers an unparalleled 22:1 overall data-reduction ratio and up to 54% lower total cost of ownership compared to direct-attached storage. And it can handle more than 400 billion files with a single namespace.
NVIDIA DGX-2 GPU servers are enabling the next generation of ML and AI applications. NetApp has published a reference architecture for using NetApp AFF storage systems with NVIDIA DGX-2 servers. The design and validation of the architecture is consistent with the NetApp Verified Architecture (NVA) for NVIDIA DGX-1 servers found in NVA-1121. For details of the validation as well as benchmark testing results that demonstrate the performance capabilities of the solution, read this report.