Fast experimentation and successful business outcomes of AI are directly correlated, but many AI projects are rife with inefficient processes. The combination of data processing time and outdated storage solutions creates bottlenecks and workload orchestration issues, and static allocation of GPU compute resources limits the number of experiments that researchers can run.
NetApp and Run:AI have partnered to simplify the orchestration of AI workloads, streamlining the process of both data pipelines and machine scheduling for deep learning (DL). With the NetApp® ONTAP® AI proven architecture, you can fully realize the promise of AI and DL by simplifying, accelerating, and integrating your data pipeline. And to help your researchers manage and optimize GPU utilization, Run:AI’s orchestration of AI workloads adds a Kubernetes-based scheduling and resource utilization platform.
Together, NetApp and Run:AI products enable numerous experiments to run in parallel on different compute nodes, with fast access to many datasets on centralized storage. With the combined solution from NetApp, NVIDIA, and Run:AI, you get an infrastructure stack that is purpose-built for enterprise AI workloads.By using Run:AI’s centralized resource pooling, queueing, and prioritization mechanisms together with NetApp ONTAP AI, your researchers are removed from infrastructure management hassles and can focus exclusively on data science. You can increase productivity by running as many workloads as you need without bottlenecks in your compute or data pipeline. And with the Run:AI scheduler and virtualization technology, you can easily use fractional GPUs, integer GPUs, and multiple nodes of GPUs for distributed training on Kubernetes. In that way, AI workloads run based on need, not on capacity. Your data science teams can run more AI experiments on the same infrastructure.
With NetApp and Run:AI technology, if your company scales AI, you get a double benefit: faster experiments and full resource utilization. To learn how to streamline and accelerate your data science initiative, read the technical report.
Mike McNamara
É líder sênior de marketing de produtos e soluções na NetApp, com mais de 25 anos de experiência em gerenciamento de dados e marketing de storage em nuvem. Antes de ingressar na NetApp há mais de dez anos, Mike trabalhou na Adaptec, Dell EMC e HPE. Mike foi um dos principais líderes da equipe que impulsionou o lançamento de uma oferta de armazenamento em nuvem de primeira empresa e a primeira solução de IA/ML conetada à nuvem (NetApp), sistema e software de armazenamento em nuvem híbrida (NetApp), iSCSI e SAS (Adaptec) e sistema de armazenamento de dados Fibre Channel (EMC CLARiiON).Além de seu papel anterior como presidente de marketing da Fibre Channel Industry Association, ele é membro do Conselho Consultivo da Conferência de Cúpula de tecnologia Ethernet, membro da Ethernet Alliance, colaborador regular de revistas da indústria e palestrante frequente de eventos. Mike também publicou um livro através da FriesenPress intitulado "Scale-out Storage - The Next Frontier in Enterprise Data Management" e foi listado como um dos 50 B2B melhores profissionais de marketing de produtos para assistir pela Kapos.Ver todas as publicações de Mike McNamara