Cognitive intelligence based 6G distributed network architecture

Abstract

5G is envisioned to guarantee high transmission rate, ultra-low latency, high reliability and massive connections. To satisfy the above requirements, the 5G architecture is designed with the properties of using service-based architecture, cloud-native oriented, adopting IT-based API interfaces and introduction of the Network Repository Function. However, with the wide commercialization of 5G network and the exploration towards 6G, the 5G architecture exposes the disadvantages of high architecture complexity, difficult inter-interface communication, low cognitive capability, bad instantaneity, and deficient intelligence. To overcome these limitations, this paper investigates 6G network architecture, and proposes a cognitive intelligence based distributed 6G network architecture. This architecture consists of a physical network layer and an intelligent decision layer. The two layers coordinate through flexible service interfaces, supporting function decoupling and joint evolution of intelligence services and network services. With the above design, the proposed 6G architecture can be updated autonomously to deal with the future unpredicted complex services.