Today we’re pleased to announce that VAST Data is officially certified for NVIDIA DGX SuperPOD, the industry’s leading accelerated infrastructure platform built for the world’s most challenging AI workloads. NVIDIA and VAST have long shared a common vision for making AI more accessible for all organizations, making today a memorable milestone in the long-standing partnership between our two companies.
As the first enterprise NAS solution certified for DGX SuperPOD, VAST introduces a new level of simplicity in high-performance storage making it easy to manage and operate even for organizations without deep HPC expertise in-house. Together, the joint offering gives customers a turnkey, co-designed, and validated “AI supercomputer” that delivers the performance needed for all modern, accelerated AI workloads.
Most in the industry would categorize VAST as an "enterprise NAS" vendor. Frankly, we’re ok with that label - the implications around operational simplicity (online upgrades and expansions for example), availability, and manageability are table stakes for an enterprise-grade product. The one stigma of enterprise NAS solutions, however, often revolves around the lack of performance for intense workloads commonly found in HPC/AI. But the world of HPC/AI is not new to VAST.
In our last fiscal year, nearly half of our business was attributed to HPC/AI workloads, with a majority of our deployments attached to GPU-equipped machines. Across a wide range of industries, our customers have experienced firsthand that - with VAST - high-performance storage doesn’t require the complexity associated with classic parallel file systems. (IDC closely examined VAST’s NFS-based approach to HPC/AI in this white paper.)
Although VAST features industry standard file and object storage protocols (like NFS and S3) similar to other legacy NAS solutions, our DASE architecture uniquely enables the VAST Data Platform to meet the demanding performance of AI workloads running on DGX SuperPOD. With NVIDIA’s certification, the world’s leading AI technology company validates what hundreds of VAST customers witness every day: a system capable of virtually limitless levels of scale and performance that’s also simple to manage and resilient to operate.
It's safe to say that interest in Generative AI and Large-Language Models (LLMs) hit a fever pitch last year with no signs of slowing down for the foreseeable future. Organizations across every industry are no longer debating whether they need to incorporate AI into their business, but rather how to incorporate AI effectively instead. Some organizations may end up simply consuming and integrating AI services (ie. chatbots, translation services) into core business applications, but for others, AI will serve as a strategic differentiator, one where they must “own their destiny.” For those in the latter group, this strategy requires new investments in people (data scientists), processes (software and tooling), but also infrastructure (compute, storage, and networking) that may be foreign to enterprise IT teams.
Ultimately, NVIDIA DGX SuperPOD with the VAST Data Platform was designed to make large-scale AI simpler, faster, and easier to manage for every organization and enterprise IT team, regardless of previous HPC experience. By democratizing AI infrastructure, VAST and NVIDIA are advancing the art of possible by integrating accelerated computing with high-performance storage to rapidly deliver insights from oceans of data. We look forward to witnessing the continued advancements in AI for years to come!
To learn more about how VAST supports simplicity in scaling AI development on NVIDIA DGX SuperPOD, join me and Tony Paikeday, Senior Director of Product Marketing, Artificial Intelligence Systems at NVIDIA, at our upcoming webinar.