PRESS RELEASE: Recently, China Mobile Research Institute, Zhejiang Mobile and ZTE jointly completed the first trial of the innovative cross-gNB wireless computing power orchestration in China. This is a further development of the innovative Radio Composer solution, which expands the computing power network in the wireless access network and lays a foundation for more AI applications on gNBs. The 5G wireless access network is an important carrier of the computing power network at the edge. China Mobile has already initiated a project of wireless computing power network research standards in China Communications Standards Association (CCSA), and is actively working with industries to carry out the research and exploration of wireless computing power network scenarios, requirements, and key technologies.

Data analysis and AI computing can help gNBs optimize radio resources and air interface processing to improve the wireless network performance and user experience. More and more data analysis and AI computing are imposing a higher requirement on hardware computing capability. At present, both the mobile communications network architecture and the software/hardware of the gNodeB are designed to meet the basic functions, real-time network performance, and stability requirements of the wireless communications system. However, the general computing resources of gNB are limited, especially in high service load scenarios, a large amount of computing resources are for gNB communication, so it is difficult to provide stable and continuous computing and storage resources for the intelligent functions of gNBs, such as data analysis and AI computing.

To break through the bottleneck of computing power of a single gNB, China Mobile Research Institute, Zhejiang Mobile and ZTE have made further exploration on the cross-gNB computing power sharing and orchestration. To build a distributed gNB computing pool by logically aggregating the computing power of multiple 5G gNBs to achieve cross-gNB computing power sharing and on-demand orchestration. The low-latency and high-efficiency meets the requirements for AI computing in terms of the Intelligent AMC (Adaptive Modulation and Coding) function of gNBs. In Jiaxing, China Mobile Research Institute, Zhejiang Mobile and ZTE have completed the first trial of cross-gNB computing power sharing and orchestration in the industry. In the scenario of unbalanced network load, the computing power in the gNB resource pool is supplemented by idle resources. With the help of gNBs with remaining computing power, gNBs can perform the data analysis and AI computing to supplement the computing power of gNBs in busy hours. In this way, the available computing power of gNBs in busy hours is increased by about 12%, and the intelligent AMC scheduling function improves the spectrum efficiency of users under medium and poor coverage by about 14%.

The innovative 5G cross-gNB computing power orchestration technology provides an important technical reference for further sharing, openness, and orchestration of wireless computing power. China Mobile will collaborate with industry partners to continuously make research and tackle the problems in terms of the wireless computing power network scenarios, requirements, and key technologies based on the wireless computing power network research standards released by CCSA. China Mobile and its partners will work on the perception, measurement, sharing, openness, addressing, collaborative scheduling, orchestration, and management & control to accelerate the construction of key capabilities of wireless computing networks. Based on the high-quality wireless computing network infrastructure, AI and computing power orchestration will be introduced to further release the potential of 5G network performance, to improve the value-added capabilities of wireless computing power, and to enhance end-to-end computing network services.