NVIDIA announced updates to Omniverse Cloud technology at its GTC event today. Applications for data center pros? Some of what NVIDIA released today seems promising for the data center world. Only Siemens and Siemens Energy appears to provide data center-centric solutions through the Omniverse Cloud platform.
Potential applications include creating a digital twin of existing data centers to forecast the impact of PUE as you expand a data center’s footprint. This technology would have a useful application in disaster planning and data recovery models based on the most likely environmental threats to data center physical infrastructure. Risk planning would be much more accurate and in a perfect world, this would reduce cybersecurity premiums in certain use cases.
At its core, today’s updates and features of Nvidia’s Omniverse open up the platform to developers. This includes Omniverse Cloud, an IaaS for securely running applications within the platform. During a press briefing yesterday, Nvidia’s vice president of Ominverse Platform Development Richard Kerris mentioned Siemens Energy is a software partner and promised more details. Our friends at Siemen’s provided us with access to a webinar on digital twin technology. In it, the software firm detailed ways data centers can:
- Twin data center infrastructure
- Simulate what-if scenarios pre-build
- Speed up the design process for multiple key performance indicators (KPIs)
This video offers an overview of the simulations built into the NVIDIA Ominverse’s Universal Scene Description (USD) offering. We're seeing a trend here in that NVIDIA wants simulation technology to be readily available and as plug-and-play as possible for enterprises.
We at Data Center Knowledge also see digital twin technology potentially removing the need for electrical workers to risk injury maintaining and upgrading highly dangerous electrical equipment, as evidenced by our recent coverage of the Google and OVH data center fires. The implementation of robotics to maintain data center equipment would get a boost from NVIDIA’s Omniverse, based on case studies from other verticals such as the automotive and railway industries.
We admit, many of the firm’s announcements have a decided cool factor, leveraging the power of 3D for realistic simulations, but is there anything here that will change your life as a data center pro today or in the very near future? Maybe. Enhancing PUE through digital twin technology by replicating power and cooling systems would be the most reachable opportunity based on what we’re seeing from Siemens Energy.
“Safety is the number one priority in robotics,” said Nvidia CEO Jensen Huang during his keynote at today’s GTC 2022 event. “All of this demands processing power.”
We in the data center industry know what that means, more demand for access to storage, network, and computer. As more industries adopt AI and digital twin technology, already tight colo space will become even more difficult to acquire. What NVIDIA is doing with this upgrade to its Ominverse platform is making AI modeling and simulations more accessible to enterprises.
To help with this, NVIDIA introduced its H100 system, calling in the “new engine of the AI factory” and it’s now available on NVIDIA LaunchPad. H100 cloud services running on AWS, Google Cloud, Microsoft Azure, and Oracle begin next year. Huang promised systems to support this leap in processing needs with semiconductor chips such as Grace Hopper, which is optimized for data analytics and recommender systems but applicable in most high-performance computing use cases. Systems to support the chips are coming in the first half of 2023.