Service-Oriented SAGIN With Pervasive Intelligence for Resource-Constrained Users

As an indispensable architecture for future 6G communication networks, space-air-ground integrated network (SAGIN) integrates satellite networks, air networks and ground networks, greatly expanding the coverage of network space. Compared with the traditional mobile edge computing (MEC), the edge intelligence (EI) formed by combining artificial intelligence (AI) and MEC can intelligently process the edge data by embedding the AI algorithms into the edge devices with limited computing power. Therefore, this article considers applying EI to SAGIN to form the EI-driven SAGIN architecture, which can significantly enhance the communication, computing, sensing and storage capabilities of SAGIN architecture to solve the problem of efficient resource management for resource-constrained users. In this article, we first introduce the system network architecture and logical functional architecture, and give a detailed description of the components in the network architecture, and then discuss some key technologies in the system, including efficient resource utilization for microservice based on software defined network (SDN) and network function virtualization (NFV), deep reinforcement learning (DRL) based on knowledge graph for efficient storage and intelligent computing, and efficient and realtime sensing for massive information. Finally, we propose a DRL-based resource allocation and computation offloading algorithm for microservices (DRCAM) and evaluate the performance of the proposed algorithm. The simulation results show that, compared with the existing algorithms, the proposed algorithm could greatly reduce the system cost under different weights.