HPE GreenLake for File Storage: Nvidia DGX BasePOD Certified and OVX Validated

By David Yu, storage product marketing, HPE

HPE GreenLake for File Storage has achieved NVIDIA DGX BasePOD certification and NVIDIA OVX storage validation, providing customers with proven enterprise file storage that accelerates AI, GenAI, and GPU-intensive workloads at scale.

GreenLake for File Storage also provides enterprise file storage for the NVIDIA AI Computing by HPE. And HPE will be a time-to-market partner on upcoming NVIDIA reference architecture storage certification programs.

With GreenLake for File Storage now certified by NVIDIA, customers can accelerate AI adoption and ROI by leveraging a fully validated reference architecture with a proven server, networking, and storage framework. This reduces the deployment and operational complexity of AI workloads, providing a fast track to AI success. By leveraging a validated storage solution for AI, customers can achieve superior performance, simplicity, efficiency, and reliability to unlock the power of AI.

NVIDIA AI computing by HPE
GreenLake for File Storage is an important component of the new NVIDIA AI Computing by HPE. The new portfolio is co-developed using NVIDIA AI computing, networking, and software, combined with HPE’s AI storage, compute, and the HPE GreenLake cloud, to deliver a full-stack, turnkey AI private cloud. The first solution is HPE Private Cloud AI, which enables enterprises of every size to gain a fast, flexible path for developing and deploying GenAI applications.

To deliver the optimal performance for the AI and data software stack, HPE Private Cloud AI provides a fully integrated AI infrastructure stack that includes the NVIDIA Spectrum-X Ethernet networking platform, GreenLake for File Storage, and ProLiant servers with NVIDIA L40 GPUs, NVIDIA H100 NVL Tensor Core GPUs, NVIDIA H200 NVL Tensor Core GPUs, the NVIDIA GH200 NVL2 platform, and NVIDIA BlueField-3 DPUs. Using NVIDIA AI Enterprise software, including NVIDIA NIM inference microservices, customers can deploy optimized AI models and applications with enterprise-grade security, support, and stability.

What is NVIDIA DGX BasePOD certification and NVIDIA OVX storage validation?
AI poses many new challenges for enterprise IT teams. Long, complex infrastructure deployment cycles can prevent quick spin-up of AI workloads that leverage company data. To overcome these challenges, NVIDIA has storage partner certification and validation programs that help enterprises pair the right storage with NVIDIA accelerated computing and high-speed NVIDIA networking.

GreenLake for File Storage has attained NVIDIA DGX BasePOD certification and NVIDIA OVX storage validation. To complete this validation, HPE ran a suite of NVIDIA tests measuring storage performance and input/out scaling across multiple parameters that represent the demanding requirements of various enterprise AI workloads. NVIDIA audits these test results to determine if the storage system passed.

Importance of storage for AI
AI is fueled by data, so having the right storage is essential to AI success. To achieve it, storage must keep up with the compute power of valuable GPU investments. Without the right storage, it’s challenging to fully utilize your GPUs because you can’t feed them with data fast enough. To keep up with the performance of GPUs, storage for AI must deliver performance at AI scale, meaning that it must be fast even for massive amounts of data.

According to IDC, 60-80% of AI projects fail, and 80% of companies plan to upgrade their storage to better support AI workloads.⁽¹⁾ Beyond performance, your storage must also deliver simplicity and efficiency at AI scale to reduce management overhead and enhance ROI, respectively.

AI involves many kinds of workloads, including computer vision, LLMs, natural language processing, and GenAI, and each goes through the various stages of AI: data aggregation, data preparation, training and tuning, and inference. The storage requirements can vary depending on the AI workloads and even the AI stages, so each customer environment needs a tailored storage solution to meet these specific needs. Each customer’s AI environment and storage needs are also unique. For example, research environments will have a greater focus on pure speed rather than enterprise features. An enterprise business environment demands an intuitive cloud experience, and data services are as important as performance. No matter where you are in your AI journey, HPE offers a rich, unique portfolio of storage offerings to address specific requirements for your AI use cases, workloads, and workflows.

GreenLake for File Storage: Storage for AI success
Although no single storage product can meet all AI needs and be the best option for every stage of AI, HPE GreenLake for File Storage spans all the stages of AI and provides a comprehensive, flexible solution for customers embarking on their AI journey. It delivers enterprise performance, simplicity, and enhanced efficiency, all at AI scale.

With this all-flash solution, you eliminate tiered AI data pipelines and avoid overprovisioning with a disaggregated shared everything architecture that allows for independent scaling of performance and capacity.

Enterprise performance at AI scale
GreenLake for File Storage is designed for fast, predictable performance to supercharge AI applications with all-NVMe drives and no front-end caching, data movement between media, or tiered data pipelines.

With support for optimized GPU utilization through NVIDIA Quantum-2 InfiniBand, NVIDIA Magnum IO GPUDirect Storage and GPUDirect RDMA, GreenLake for File Storage accelerates AI workloads by boosting performance for model training and tuning via faster checkpointing.

Simplicity at AI scale
It provides an intuitive cloud experience with radically simple file data management that reduces operational and management overhead for your IT staff and increases productivity for your data scientists and line-of-business application owners. Powered by HPE GreenLake cloud, you get streamlined deployment, simple setup, easy file share creation, fast job completion, automated, non-disruptive upgrades, and unified storage management, all from a single cloud console accessible from anywhere, on any device.

Efficiency at AI scale
Data-intensive AI models have a need for storage capacity, rack space, and power, so efficiency is crucial. GreenLake for File Storage can reduce your AI storage costs with high-density capacity options with low power consumption – as well as further lower your energy consumption with industry-leading data reduction, non-disruptive upgrades, and an efficient AI storage as-a-service consumption model. With more efficient GPU utilization, you can also get a higher GPU ROI along with enterprise performance at AI scale.

Another step in HPE’s AI journey with NVIDIA
With NVIDIA DGX BasePOD certification and NVIDIA OVX storage validation for GreenLake for File Storage, HPE has taken another major milestone in our AI journey with NVIDIA. This is part of HPE’s commitment and trajectory to deliver real-world solutions through the power of AI.

Resources :
HPE GreenLake for File Storage blog series
Building a modern AI stack with HPE GreenLake for File Storage solution brief
HPE GreenLake for File Storage architecture technical white paper
Technical overview: Inside HPE GreenLake for File Storage technical blog
Video: HPE GreenLake for File Storage technical demo
Video: HPE GreenLake for File Storage: Accelerate and Simplify AI Workloads with AI-ready Storage