Everpure FlashBlade NVIDIA Cloud Partners Reference Architecture

NVIDIA

Audience
Public
Product
FlashBlade
Content Type
User Guides
Technology Integrations
NVIDIA
Source Type
Documentation

Executive Overview

The Everpure® FlashBlade//S500 is a certified, high-performance, unified storage platform tailored for NVIDIA Cloud Partners (NCPs) building infrastructure for AI workloads on NVIDIA HGX™ servers.

FlashBlade® delivers the performance, operational efficiency, scalability, and simplicity required to support the rapid evolution of GPU-accelerated workloads across training, inference, and data preprocessing pipelines. By integrating seamlessly into the NVIDIA Cloud Partner ecosystem, FlashBlade enables NCPs to build scalable AI-as-a-Service environments with consistent performance and secure multitenancy.

This document presents a validated reference architecture for NCPs that combines Everpure FlashBlade//S500 and Everpure Fusion™, a federated FlashBlade//S500 fleet management solution, with NVIDIA HGX H100, H200, and B200 platforms. The solution is optimized to deliver predictable, low-latency data access across thousands of GPUs and supports the high-throughput demands of modern AI pipelines and adheres to the NCP Reference Architecture (NCP RA) specifications for high-performance storage (HPS)

Designed for unstructured data at scale, FlashBlade//S500 helps NCPs accelerate time-to-value while reducing infrastructure complexity and power consumption. The rest of this document details how FlashBlade//S500 meets the NCP HPS requirements, with a focus on deployment, scalability, and multitenant support.