2 months ago
24306
ARTICLE AD BOX
Felix Pinkston Sep 19, 2024 01:32
NVIDIA launches AI Aerial to optimize wireless networks and deliver generative AI experiences, collaborating with key telecom players like T-Mobile, Ericsson, and Nokia.
Telecommunications providers are advancing beyond traditional voice and data services with the introduction of NVIDIA AI Aerial, an AI computing infrastructure designed to optimize wireless networks and cater to the next-generation needs of generative AI. According to the NVIDIA Blog, this platform, launched on September 18, 2024, represents a significant milestone in AI and telecommunications convergence.
Driving Networks of the Future With AI-RAN
NVIDIA AI Aerial is the world’s first AI-RAN (Radio Access Network) platform that can host generative AI and RAN traffic while integrating AI into network optimization. AI-RAN offers high-performance, energy-efficient, software-defined RAN, enhancing network experiences and creating new revenue opportunities through edge AI applications for both internal and third-party generative AI solutions.
This technology is foundational for the multipurpose networks of tomorrow, leveraging AI-powered telecommunications capabilities to meet emerging demands.
Harnessing NVIDIA AI Aerial for Telecom Industry
The NVIDIA AI Aerial platform provides a comprehensive suite of capabilities, including high-performance, software-defined RAN, training, simulation, and inference. This allows telecom operators to engage at any stage of development to deployment for next-generation wireless networks.
Key components of the NVIDIA AI Aerial platform include:
- NVIDIA Aerial CUDA-Accelerated RAN: Software libraries enabling high-performance virtualized RAN workloads on NVIDIA-accelerated compute platforms.
- NVIDIA Aerial AI Radio Frameworks: PyTorch- and TensorFlow-based software libraries for developing and training models to improve spectral efficiency and enhance 5G and 6G radio signal processing. This also includes NVIDIA Sionna, a simulator for neural network-based radio algorithms.
- NVIDIA Aerial Omniverse Digital Twin (AODT): A system-level network digital twin development platform enabling realistic simulations of wireless systems, integrating software-defined RAN and user-equipment simulators.
NVIDIA AI Aerial and AI-RAN Innovation Center
NVIDIA is partnering with T-Mobile, Ericsson, and Nokia to accelerate AI-RAN commercialization through the AI-RAN Innovation Center. This collaboration aims to merge RAN and AI innovations, delivering transformative network experiences for customers.
Tommi Uitto, president of Mobile Networks at Nokia, remarked, “AI-RAN is set to revolutionize the telecom industry, and the AI-RAN Innovation Center will drive industry collaboration. By uniting leading companies in telecom and AI, we can unlock AI’s full potential in our networks, improving performance, reducing costs, and creating new opportunities for our customers.”
Fredrik Jejdling, EVP and head of Business Area Networks at Ericsson, added, “Ericsson has invested in our AI-RAN solution, enabling communications service providers to deploy portable RAN software across multiple platforms. We are now evaluating NVIDIA accelerated computing’s performance and cost in this context.”
NVIDIA AI Aerial Ecosystem
The growing ecosystem of NVIDIA AI Aerial includes key partners like Softbank and Fujitsu. Ansys and Keysight are utilizing NVIDIA Aerial Omniverse Digital Twin for testing and simulation, while institutions like Deepsig, ETH-Zurich, Northeastern University, and Samsung collaborate on 6G research and NVIDIA Aerial AI Radio Frameworks.
Cloud stack software providers such as Aarna Networks, Canonical, Red Hat, and Wind River, along with networking stack providers like Arrcus and server infrastructure providers like Dell Technologies, Hewlett Packard Enterprise, and Supermicro, are significant partners. Edge solution providers like Vapor.io and system integrators like World Wide Technology, with its AI Proving Ground, are accelerating AI solution decision-making.
Learn more about NVIDIA AI Aerial
Image source: Shutterstock
English (US) ·