Accelerating 5G Service-Based Architecture with Ebpf
Yong-Xuan Huang, Kung Chun Wang, Bi Shun Ke · 2023
Service-Based Architecture (SBA) serves as a foundational methodology for 5G core networks, as delineated by the 3rd Generation Partnership Project (3GPP). This modular approach segments the 5G Core Network into discrete, functionally autonomous Network Functions (NFs), deployed via either virtual machines or container technologies. The efficacy of the SBA execution platform significantly impacts the service capabilities of these NFs. In this study, we present an enhanced architecture that employs extended Berkeley Packet Filter (eBPF) and eXpress Data Path (XDP) technologies to fine-tune the network, security, service, and scheduling aspects within a Kubernetes framework. Additionally, eBPF’s monitoring and packet control features augment the service capabilities of 5G components, thereby improving encrypted data transmission. eBPF’s real-time monitoring aids the scheduler in making nuanced decisions, while an eBPF-driven load balancer ensures equitable service distribution. Comprehensive testing reveals that the enhanced architecture increases transmission efficiency by 17%, reduces average latency by 43%, and alleviates latency anomalies compared to conventional 5G NF platforms. It also enhances the inherent connectivity of 5G NFs by approximately 15%. For encryption-reliant applications, the architecture achieves a 17% performance gain in latency, transmission, and connectivity, while its scheduling and service balancing capabilities demonstrate a roughly 19% performance enhancement.