Hyperscale data center operators are poised to invest billions to build out AI infrastructure, and this has placed a spotlight on the role of software-defined storage in new platform deployment.

For Supermicro Computer Inc., this represents an opportunity to not only showcase its portfolio of servers, storage systems and switches, but also to work with its extensive partner network in bringing innovative storage solutions to market.

“Supermicro comes into play from the platform and infrastructure level,” said Vince Chen, director of solutions architecture at Supermicro. “Through the work and then the business [with] customers and partners, we actually are able to provide our customers the infrastructure level of capabilities and offerings. We call it data center building block solutions. Together, we are able to serve customers at any scale from the edge, enterprise, on-prem, all the way to the larger scale, like AI or GPU data centers.”

Chen spoke with theCUBE’s Rob Strechay at the Supermicro Open Storage Summit, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. He was joined by Sherry Lin, storage solution product manager at Supermicro; William Li, director of solution management at Supermicro; John Fragalla, principal architect at Nvidia Corp.; Balaji Venkateshwaran, vice president of AI product management at DataDirect Networks Inc.; Steve Umbehocker, founder and chief executive officer of OSNexus Corp.; and Jeffrey Slapp, chief technology officer of SteelDome Cyber LLC. They discussed how organizations leverage various technologies to provide cutting-edge storage platforms for cloud-native, AI and traditional applications. (* Disclosure below.)

Multi-tenant support in software-defined

One element of Nvidia’s partnership with Supermicro involves the Nvidia Cloud Partner program. A key factor for Nvidia and the companies that participate in its program is the ability to support multiple tenants coming from different organizations to share NCP infrastructure. This is important for GPU resource allocation, according to Nvidia’s Fragalla.

“Multi-tenant is allowing them to be assigned to a certain number of GPUs for their quality-of-service settings to ensure that they could do the amount of specific workloads and the amount of GPUs they need,” he explained. “It comes down to the storage must support multi-tenant and quality of service settings to ensure…enough bandwidth and the amount [of] capacity for each tenant.”

Nvidia also operates a certification program in which storage partner platforms are analyzed according to strict performance criteria, including factors such as scale, reliability and data integrity features. DataDirect Networks recently completed Nvidia’s certification program for its AI400X3 scalable data storage offering as part of its EXAScaler platform.

“With the partnership of Nvidia and Supermicro, we at DDN have had the privilege to service a number of our AI customers at scale through the high-performance DDN EXAScaler storage capabilities,” Venkateshwaran said. “The throughput of tokens, time to first token, the performance and low latency characteristics … DDN EXAScaler specifically has had a multi-generation proven track record in this. We’ve been able to service a number of happy customers at scale, including GPU deployments as large as 100,000-plus GPUs.”

Petascale all-flash for on-prem storage

Supermicro and its partner OSNexus have collaborated on a solution that brings a single point of management for file, block and object storage. This has led to implementations using large scale software-defined storage deployments that include Supermicro’s Petascale all-flash array with the Nvidia Grace CPU superchip.

“We have very cost-effective on-prem storage solutions for large scale enterprise and service providers,” Li said. “With OSNexus, we do have a lot of success stories. The top-notch one would be the Petascale all-flash families along with all-hyperscale all-flash families.”

Where are these cutting-edge high-performance storage solutions being deployed? Academic institutions have been a central focus, according to OSNexus’ Umbehocker.

“The largest growth in 2024 has been in higher education,” he said. “There’s been some very large scale-out clusters deployed around the Supermicro platforms. We’ve also made a lot of strides in some of these other verticals … around financial services, government and healthcare.”

Leveraging server and node system solutions

Supermicro’s partnership with SteelDome is designed to deliver validated solution-ready infrastructures that are tailored for different use cases. The two firms provide enterprise-grade, hyperconverged infrastructure that leverages Supermicro’s server and node system technologies.

“We have a diverse platform,” Lin said. “The user depends on their use case from compact models like the IoT SuperServers, for edge locations … or to the rack space friendly multi-node system like BigTwin, which is ideal for the compute-intensive workload.”

The collaborative hyperconverged infrastructure approach also integrates SteelDome’s StratiSystem, a modular data-centric operating system that can power a suite of solutions. Operational ease-of-use is a key goal here, according to Slapp.

“The platform’s modular architecture is designed for simplicity, making it easy to design, deploy and adapt to the widest range of infrastructure requirements,” he noted. “The platform is designed to run seamlessly on any Supermicro server or storage system today and into the future, ensuring both compatibility and long-term flexibility as hardware evolves.”

Here’s a short clip from our interview, part of SiliconANGLE’s and theCUBE’s coverage of the Supermicro Open Storage Summit:

(* Disclosure: TheCUBE is a paid media partner for the Supermicro Open Storage Summit. Neither Super Micro Computer Inc., the sponsor of theCUBE’s event coverage, nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE