Liquid Software-Based Edge Intelligence for Future 6G Networks

• Published in: IEEE network Vol. 36; no. 1; pp. 69 - 75
• Main Authors: Yang, Tingting, Qin, Meng, Cheng, Nan, Xu, Wenchao, Zhao, Lian
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.01.2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: 6G mobile communication; Artificial intelligence; Big Data; Computer architecture; Data models; Deep learning; Edge computing; Electronic devices; Energy storage; Intelligence; Intelligent networks; Modular design; Modular equipment; Network design; Software; Software-defined networking; Training data; Wireless communication; Wireless networks
• ISSN: 0890-8044; 1558-156X
• DOI: 10.1109/MNET.011.2000654

The 6G wireless network is promising to build bridges toward smart society in the digital world, which calls for innovative architectures and new solutions. The future 6G network should be sensing-based and data-driven for near-instant and massive connectivity with distributed intelligence. With a majority of intelligent applications being deployed at the edge, artificial intelligence (AI) is envisioned to play a key role in satisfying key requirements of 6G networks. Edge intelligence, as the marriage of AI and edge computing, is envisioned to fully meet the potential requirements of edge big data with energy, bandwidth, storage, and privacy concerns. However, it is an attractive issue to deal with distributed edge intelligence for the complexities and heterogeneous requirements, especially considering the time-varying channels and network dynamics. Furthermore, the ever increasing number of smart devices present great challenges for intelligent network management and newly modular network design in 6G networks, which needs to enable liquid self-management with comprehensive network intelligence. Hence, in this article, we first comprehensively give an overview on AI toward 6G networks, and characterize the requirements of a 6G network for AI applications. In particular, we investigate distributed edge intelligence challenges, requirements, and trends in future 6G networks. Then a liquid-specific and flexible software-defined network architecture for AI applications is inspired and discussed by 6G networks, which will play a crucial role in both academia and industry.