There are over 7 million data centers worldwide. IT organizations are adopting the latest technologies to handle exponentially increasing data and reduce costs.

NVIDIA is streamlining the collaborative virtual design process with its NVIDIA Omniverse simulation platform to help build digital twins that accurately represent every component of an AI supercomputing facility.

NVIDIA announced on the 15th that it will showcase the NVIDIA Omniverse demo at SC22. The demo shows how to improve the design and development of complex supercomputing facilities using an open development platform behind the data center.

'Omniverse' enables data center operators to aggregate real-time data inputs from key third-party computer-aided design, simulation and monitoring applications, enabling them to view and work with their entire data set in real time.

This demo demonstrates how Omniverse enables users to leverage accelerated computing, simulation, and operational digital twins connected to real-time monitoring and AI. This is expected to simplify facility design, accelerate construction and deployment, and optimize ongoing operations.

We'll also be showing the 'NVIDIA Air Demo'. NVIDIA Air is a data center simulation platform that works with Omniverse and is designed to simulate the network, the central nervous system of a data center.

NVIDIA AIR enables you to model the entire network stack, automating and validating network hardware and software before going live.

'Patch Manager' is an enterprise software application that plans point-to-point connections at the cabling, asset, and physical layers across a network domain.

When Patch Manager is connected to Omniverse, complex topologies of port-to-port connections, rack and node layouts, and cabling can be integrated directly into the live model. This allows data center engineers to see the dependencies and overall picture of the model.

Additionally, NVIDIA Air enables simulation and pre-validation of accurate network topologies, including protocols, monitoring, and automation.

Once the machine learning data center is built, sensors, control systems, and telemetry can be connected to the Omniverse internal digital twin to monitor operational status in real time.

Digital twins can help engineers simulate common risk events, such as power peaks or cooling system failures, to improve energy efficiency.

Digital twins can also be used to test and validate software and component upgrades before they are deployed in physical data centers.