Just days after sealing a deal with Ericsson to explore converged architectures for artificial intelligence (AI)-based radio access networks (RAN), Japanese IT business and comms operator SoftBank has announced a collaboration with Red Hat on the joint research and development of AI-RAN technologies.
SoftBank is among the companies that have formed the AI-RAN Alliance, a collaborative effort to integrate AI into cellular technology to progress RAN technology and mobile networks. This encompasses work to explore common network and compute infrastructure services that allow AI and RAN to operate on the same network infrastructure, with the ultimate aim being the ability to unlock new use cases for communications service providers by using the power of AI to enhance network efficiency.
With Ericsson, SoftBank is also conducting techno-economic analyses, developing prototypes and running lab demonstrations to optimise RAN and AI convergence at the edge.
In this latest development, RedHat and SoftBank will look to develop solutions intended to help maximise the use of AI for enhanced network orchestration and optimisation, and aim to deliver intelligent, autonomous networks with improved performance, higher resource efficiency and a better user experience. These are said to be particularly critical in supporting 5G and future 6G network use cases where AI-enabled applications can use these solutions for differentiated service offerings and new revenue streams.
Three key technology areas are under investigation: support for AI and RAN functions on ARM architecture; performance optimisation of distributed units (DU) on hardware-accelerated ARM architecture; and development of an orchestrator.
Red Hat and SoftBank are also looking to advance the integration of high-performing AI technology, running on Red Hat OpenShift, within the RAN infrastructure to enable faster data processing and resource optimisation to achieve enhanced performance, intelligent automation and enhanced multi-layer security.
To make this technology available to the broader telecoms industry, SoftBank and Red Hat plan to work together to help deliver new AI use cases through the development of an open and interoperable RAN ecosystem.
“AI has the potential to revolutionise how service providers operate, and a standardised platform is crucial for cost-effective, efficient and automated operations. Through collaborations like this one, Red Hat and SoftBank are able to help service providers navigate this terrain by making AI easier to develop and deploy from the edge to core to hybrid cloud and back,” said Chris Wright, senior vice-president of global engineering and chief technology officer at Red Hat.
“This joint effort will help service providers leverage AI models in the RAN for streamlined operations, new and exciting applications and more automation than ever before,” he added.
Ryuji Wakikawa, vice-president and head of the research institute of advanced technology at SoftBank, said: “AI-RAN’s software foundation serves as a robust RAN platform that not only meets the stringent performance requirements of RAN, but also enables seamless AI application integration. We have newly designed and developed the orchestrator specific to the AI-RAN, showcasing our ability to operate this essential component. Our collaboration with Red Hat will accelerate the ongoing evolution … ensuring rapid advancements.”