The NVIDIA GPU Technology Conference (GTC), the premier artificial intelligence (AI) and deep learning conference of the year for developers, has shifted to an online event—GTC Digital—due to growing concern over the coronavirus pandemic. NetApp is supporting the digital event and is announcing solutions that complement NVIDIA’s AI efforts, including a unique AI control plane that simplifies data-pipeline and provisioning tasks for data scientists and engineers, as well as the latest innovations for vertical markets.
Enterprises across all industries view AI as essential for staying competitive in a digital economy, requiring big changes to the enterprise data center. A recent IDC white paper, NetApp Solutions for AI Drive Better Business Outcomes1, reported that just 18% of organizations had AI models in production, 16% were in the proof-of-concept stage, and 15% were experimenting with AI.
NetApp and NVIDIA are delivering AI solutions that help enterprises accelerate AI adoption and manage projects more easily as they progress from pilot to production. NetApp’s attention to optimizing data pipelines amplifies the rapidly expanding ecosystem of NVIDIA AI hardware and software.
“NetApp is an important partner for NVIDIA as we work to develop solutions that meet enterprise AI needs,” says Charlie Boyle, VP and GM of DGX systems. “NetApp storage solutions and NetApp AI Control Plane are a perfect complement to NVIDIA DGX systems for rapidly evolving enterprise AI environments.”
Empower Data Scientists and Engineers with the NetApp AI Control Plane
The NetApp AI Control Plane addresses the AI data management needs of the enterprise. With the AI Control Plane, data scientists no longer have to wait for copies of datasets, and your organization no longer has to dedicate so much costly high-performance storage to store many copies of the same data. The headaches and risks that result from trying to track changes across multiple different versions of the same dataset simply disappear.
“Data scientists and data engineers are struggling to manage AI implementation details today,” says Ritu Jyoti, Program Vice President, Artificial Intelligence Research and Global AI Research Lead, IDC. “The types of capabilities offered by the NetApp AI Control Plane can help simplify AI operations. Demand for these capabilities is growing as enterprises develop more AI projects. Vendors that offer an ecosystem of solutions can help end users simplify the AI environment and avoid piecemeal solutions.”
Using open source technologies including Kubernetes, Kubeflow, and NetApp Trident, the AI Control Plane automates data pipeline operations for data prep, training, validation, dev/test, and deployment. With plug-and-play workflow automation, your teams can perform AI training runs with automatic traceability and versioning. Exact copies of production data can be cloned in seconds, with no fear of interfering with production.
Instant data accessibility facilitates rapid experimentation. Data scientists can provision Jupyter notebooks that have the ability to access petabytes of data while managing datasets, model versioning, and baselining from within Jupyter.
Built-in versioning enables full dataset-to-model traceability with seamless switching to support dev/test, A/B testing, and other needs. Workspaces and workloads can span edge, core, and cloud locations, unifying AI compute and data silos across sites and regions, simplifying operations and enabling cross-site collaboration.
Speakers: Mike Oglesby (NetApp) and Karthik Nagalingam (NetApp)
Gain a fuller understanding of the capabilities of the NetApp AI Control Plane’s, including the ability to:
- Clone a data namespace as easily as a Git repository
- Create AI data and model baselines for traceability and versioning
- Provision Jupyter workspaces with access to full datasets
- Aggregate and unify AI compute and data silos across sites and regions
Helping Enterprises Build Their AI Center of Excellence
Maintaining separate silos of AI infrastructure to address the needs of different teams or projects creates complexity, drives up costs, and constrains performance. NetApp and NVIDIA are creating enterprise-scale solutions that eliminate infrastructure silos and make your AI teams more efficient:
- NetApp® ONTAP® AI. This market-leading solution—combining NVIDIA DGX systems and NetApp AFF storage—was introduced in 2018 and has been deployed at 65 customer sites across the globe in the last 14 months.
- NVIDIA DGX SuperPOD™. NetApp joined with NVIDIA to introduce DGX SuperPOD solutions in November at SC19. Full details on the architecture are now available here.
Both NetApp and NVIDIA promote the idea of an AI Center of Excellence, an IT-led AI infrastructure platform accessible across the company that brings together people, processes, and technology. This approach yields significant benefits as illustrated in the following figure, and enables your teams to iterate faster, automate reproducibility, and deliver AI projects up to three months sooner and with higher quality.
DGX-Ready Data Centers and AI Anywhere
NetApp and NVIDIA have partnered with ScaleMatrix and DDC to offer convenient, end-to-end solutions that deliver the power and cooling required by high-performance infrastructure.
“AI is driving amazing advances in nearly every industry. Being able to support the increase in power and cooling within traditional data centers, and the ever-growing number of edge locations now needing advanced infrastructure solutions is a real challenge,” said Chris Orlando, CEO and Co-Founder at ScaleMatrix. “By working closely with NetApp and NVIDIA, we are able to deliver energy efficient, scalable, and cost-effective solutions for deploying IT resources in any environment, while new reference architectures delivered via the AI Anywhere ecosystem help simplify some of today’s most challenging AI workloads.”
With NetApp, NVIDIA, and ScaleMatrix powered by DDC, you can truly deploy supercomputing capabilities and AI anywhere.
You can learn more about building an AI Center of Excellence in the following GTC Digital on-demand session:
Speakers: Santosh Rao (NetApp) and Tony Paikeday (NVIDIA)
Uncover the secrets for operationalizing AI across the enterprise by building a mature service that enables data science teams to build AI applications faster.
Medical Imaging Solution for Healthcare
NetApp is the first storage vendor to bring to market an integrated medical imaging solution based on NVIDIA Clara Train SDK v2.0 and leveraging NVIDIA DGX-2™ systems and NetApp AFF storage. As shown in the figure below, the complete solution demonstrates a data pipeline architecture for MR image acquisition, data curation, AI-assisted annotation, and transfer learning to fine-tune AI models.
For this solution, NetApp worked closely with the NVIDIA Clara team, with the resulting report fully reviewed and approved by NVIDIA. For full information, read the NetApp technical report, ONTAP AI Reference Architecture for Healthcare: Diagnostic Imaging.
You can learn more about this and other industry solutions by viewing the following GTC Digital on‑demand session:
Speakers: Dave Arnette (NetApp) and Rick Huang (NetApp)
Learn about NetApp AI solutions for autonomous vehicles, healthcare, financial services, and retail use cases. We will discuss specific models, data, workflows, and necessary infrastructure for each use case, and explain how NetApp reference architectures streamline deployment, operational management, and software development workflows using native integration with industry-standard IT and data science tools.
More Information and Resources
To learn more about the full range of NetApp AI solutions, visit netapp.com/ai.
Check out these additional resources to learn more about solutions described in this blog and other NetApp AI solutions.
- NetApp AI Control Plane solution brief
- NetApp EF600 All-Flash Array with NVIDIA DGX SuperPOD solution brief
- NetApp ONTAP and Lenovo for Entry-Level AI/ML solution brief
- Operationalize Data Science at Scale with NetApp and Iguazio solution brief
1 IDC white paper sponsored by NetApp: NetApp Solutions for AI Drive Business Outcomes Across Core and Cloud, November 2019