Load Balancing in Connected Vehicle Services via Dynamic Edge Server Selection with Telco API

With the increasing number of connected vehicles and the diversification of services, coupled with the expanding use of AI, the demand for computing resources and communication capabilities continues to grow. To address these demands, edge servers are expected to reduce latency and enhance communication efficiency. However, edge servers generally possess limited computing resources compared to cloud servers, raising concerns about service degradation due to overload on specific servers. Consequently, service providers must efficiently distribute traffic across edge servers based on their operational status and application requirements. In mobile networks beyond 5G, methods have been explored to enable service providers to dynamically adjust network policies through Telco APIs, facilitating real-time adaptation to service demands and enhancing flexibility. Telco APIs serve as an interface between mobile networks and external systems, allowing immediate reflection of service provider requests within the network. This study proposes a method for transmitting traffic distribution requests from service providers to the mobile network via Telco APIs, enabling the optimal selection of the User Plane Function (UPF) to achieve traffic load balancing among edge servers. We implement the Network Exposure Function (NEF) and Telco API using the open-source software Free5GC on geographically distributed edge servers. By dynamically modifying routing decisions within the mobile network in response to service provider requests, we demonstrate that this approach mitigates edge server overload, prevents service quality degradation, and supports more flexible and stable vehicle-oriented applications.