NVIDIA recently announced that it was creating the first AI-on-5G Open Innovation Lab in partnership with Google Cloud. It is intended to assist industry players and AI software partners with developing, testing, and adopting solutions to help improve city infrastructure, factories, and other 5G and AI applications.
The labs will enable enterprises to use the Anthos platform from Google Cloud and NVIDIA’s accelerated processing hardware and software platforms to process data and implement AI. This would allow enterprises to increase their performance, cut costs, boost safety and reliability.
What is Anthos Platform?
Anthos is a managed application platform that extends Google Cloud services and engineering practices to the environment so that users can modernise applications faster and achieve consistency in operation.
Key features of Anthos:
- Anthos enables users to run Kubernetes clusters in cloud or on-premises environments.
- It integrates security into each stage of the application life cycle– from development to operation. In doing so, Anthos can use a defence-in-depth security strategy that incorporates comprehensive controls spanning all application deployment models.
- Anthos Service Mesh helps operations and development teams to manage and safeguard traffic across services during the monitoring, troubleshooting and improved application performance.
- Authors can run without a hypervisor layer on existing virtualised infrastructure and bare-metal systems.
Furthermore, Anthos supports NVIDIA GPU-accelerated servers, providing a uniform deployment and operational experience across deployments while lowering costs and increasing developer productivity. Additionally, the platform safeguards applications and software supply chains. Finally, it takes an outcome-driven approach to manage policies across environments – a vital feature for network operators and organisations deploying AI on 5G.
Developing AI-on-5G ecosystem
According to NVIDIA’s Senior Vice President of Telecom- Ronnie Vasishta, the global telecommunications industry is undergoing a sweeping transformation as the speed, and low latency of 5G creates unprecedented opportunities to deliver AI-on-5G at the enterprise edge. With Google Cloud as its collaborator, operators and infrastructure players will have a chance to envision and construct new profit centres built on AI and machine learning.
In April, NVIDIA announced that it was teaming with Google Cloud, Fujitsu, Mavenir, Radisys and Wind River to build solutions for NVIDIA’s AI-on-5G platform. To present a framework for both enterprises and telecom providers, Google Cloud is extending the Anthos application platform to the network edge, allowing service providers and companies to rapidly build new services and applications at the 5G edge. Enterprises can rely on Google Cloud managed services and NVIDIA for their IoT economy and harness data and AI to increase corporate performance, operational efficiency, safety, and reliability.
The AI-on-5G platform built on the NVIDIA Aerial software development kit and the NVIDIA BlueField-2 A100 convergence card combines GPUs and DPUs and incorporates NVIDIA’s ‘5T for 5G’ solution. This dramatically boosts the efficiency of 5G RAN and AI systems, resulting in innovative projects like guided vehicles, robotics, drones, and aisles that serve hundreds of other life-changing applications.
AI-on-5G data centers
Software-defined RANs are crucial for constructing a new 5G infrastructure that can operate on a shared platform. NVIDIA Aerial allows the possibility by providing elasticity and flexibility to offer services according to the dynamic customer needs as network traffic fluctuates throughout the day. The NVIDIA EGX platform, provided by OEMs using NVIDIA-Certified Systems, leads to AI computational capabilities. Server builders can also build hyper-converged edge data centres on a platform by pairing NVIDIA GPUs and DPUs.
The NVIDIA Aerial SDK ( in combination with the NVIDIA Metropolis for smart cities, NVIDIA Isaac for autonomous machines, and NVIDIA Clara for healthcare) is a critical component of the AI-on-5G ecosystem. It can be easily deployed on a single NVIDIA-Certified System.
NVIDIA revealed that the NVIDIA BlueField-3 A100 would integrate 16 Arm-based CPU cores as part of its next-generation Aerial A100 AI-on-5G computing platform. This results in a self-contained, convergent card that enables the delivery of enterprise edge AI applications across cloud-native 5G vRANs with increased performance per watt and faster implementation.
While these solutions will be validated and tested in the AI-on-5G lab, the companies will begin development in the second half of the year.