Dynamic Energy-Efficient User Plane Function Framework

This paper presents a technical report on a framework designed for the dynamic selection of User Plane Function (UPF) implementations within 5G networks, focusing on energy efficiency. The UPF is identified as a major component in the 5G Core, significantly contributing to energy consumption due to static deployment configurations that do not adapt to varying traffic loads. The proposed framework utilizes traffic prediction through a Long Short-Term Memory (LSTM) model to optimize energy usage while maintaining service performance. It profiles two distinct UPF implementations, assessing their power consumption and performance under different traffic conditions. A utility function is introduced to guide the selection process based on predicted performance and energy efficiency. The report details the methodology, including profiling, traffic forecasting, and decision-making logic, and presents simulation results demonstrating a potential energy saving of approximately 10.46% compared to static configurations. The findings suggest that intelligent management of UPF resources can enhance the sustainability of 5G network operations.