On the Automated Scaling of User Plane Function for 5G: An Experimental Evaluation

The advent of emerging technologies like autonomous vehicles, AR/VR, IoT, smart cities, etc., have raised the bar in terms of dealing with enormous data transfer, high speeds, and low response times. 5G has been designed and is optimized unceasingly to try to meet these demands. The quick evolution of 5G Radio Access Network (RAN) technologies has ushered in a new era of connectivity and communication. This has caused the 5G Core Network to face significant problems, as on many occasions, the RAN’s ability to transmit data overwhelms the core network’s ability to handle it. As a result, network congestion occurs, leading to reduced network speeds, significant delays, and occasional disruptions, which have a profound effect on low-quality user experience. This issue has triggered the scientific community to investigate ways to enhance the performance of the 5G core network, to match with the evolving demands of RAN technologies. In this work, we present an innovative algorithm for dynamic management and optimization of 5G network resources within a Kubernetes cluster environment. The algorithm’s functionality revolves around monitoring metrics of User Plane Function (UPF) and making real-time decisions on deploying multiple UPFs within the cluster to ensure enhanced network performance and cost optimization.