Performance Evaluation of Congestion Control Over B5G/6G Fluctuating Scenarios

In recent years, Internet traffic has grown constantly, and new high-throughput applications have been introduced and widely used by mobile users. This has necessitated improvements in the access networks and transport protocols. Beyond 5G and 6G (B5G/6G), cellular systems play a central role in providing low latency and high bandwidth through a new spectrum, particularly with millimeter waves (mmWaves). This technology opens up new scenarios for development, but it is necessary to determine whether transport protocols that incorporate congestion control will fully utilize the available capacity and if their performance meets the requirements of the applications. In this study, we provide a comparison of state-of-the-art algorithms for congestion control in highly dynamic B5G/6G mobile scenarios. In addition to standard metrics such as throughput and delay, we propose the convergence time criterium to determine the algorithms' response to changing conditions and to variable capacity due to mobility, signal obstructions, and high-fluctuating signal levels in the mmWave band. Simulations were conducted on 5G-LENA with different combinations of congestion control algorithms, edge vs. remote server deployments, radio link control buffer sizes, and frequency ranges. Results indicate that HighSpeed and New Reno are the slowest to converge, whereas BBR is the fastest.