D5RAN: A Substrate Platform for Service-Based RANs Toward the Integration of Computing and Communication
Zecheng Fang, Chunjing Yuan, Lin Tian, Na Li, Shuyuan Zhang, Zhou Tong · IEEE Communications Magazine · 2025
The evolution toward 6G mobile networks will extend 5G capabilities through integrated communication-computing architectures, creating intelligent wireless networks. As fundamental components of radio access networks (RANs), base stations require enhanced capabilities to support AI-driven applications. However, enabling this communication for the AI paradigm presents significant challenges, such as optimizing resource allocation and ensuring reliability in less robust cloud environments. This article proposes D5RAN — a substrate platform designed for the service-based RAN architecture that enables efficient communication-computing integration. Through five-dimensional decoupling in service development and operations (DevOps), D5RAN facilitates adaptive RAN protocol stack configuration and reliable data delivery. We analyze implementation challenges and feasible architectures of service-based RANs for AI, presenting D5RAN's framework. A proof-of-concept based on D5RAN demonstrates dynamic communication service orchestration for AI demands, highlighting the potential of service-based RANs in advancing communication for AI within 6G networks.