The relentless demand for faster data rates and reduced end-to-end delays has driven the evolution of communication networks towards technologies like 5G and beyond [3, 5, 6, 11, 18, 19]. High-speed, low-latency applications, ranging from real-time gaming and augmented reality to autonomous systems and industrial automation, are critically dependent on efficient and predictable network performance. Traditional congestion control mechanisms, primarily embodied in variants of the Transmission Control Protocol (TCP), often react to congestion signals (like packet loss or delay) rather than predicting them [1, 4, 7, 8, 9, 10, 12, 13, 14, 15, 16, 17, 20, 21]. While various TCP variants have been developed for different network conditions [7, 8, 9, 10, 12, 13, 16, 17, 19, 20, 21] and machine learning has been explored for congestion control [5, 6, 18], a proactive approach through accurate prediction is highly desirable for meeting stringent low-latency requirements. Rough Set Theory (RST) [24, 25] is a mathematical framework for dealing with vagueness and uncertainty in data, offering powerful tools for attribute reduction and rule extraction [24, 25]. Although RST has been applied successfully in various prediction and classification tasks in other domains [24, 25], its application to predicting network congestion, particularly from the perspective of network towers supporting high-speed, low-latency traffic, is a novel area. This article proposes a design concept for a congestion prediction model leveraging RST, outlining its methodology, expected benefits in terms of interpretability and efficiency, and potential for deployment on network infrastructure to enable proactive congestion management and enhance high-speed, low-latency communication.

Rough Set Theory, Congestion Prediction, High-Speed Communication, Low-Latency Networks, Network Optimization

L. Li, P. Sharma, M. Gheisari and A. Sharma, “Research on TCP performance model and transport agent architecture in broadband wireless network”, Scalable Computing: Practice and Experience, 2021, vol. 22, no. 2, pp. 193-201.

M. Prakash and A. Abdrabou, “On the fidelity of ns-3 simulations of wireless multipath tcp connections”, Sensors, 2020, vol. 20, no. 24, pp. 7289.

M. H. Hanin, M. Amnai and Y. Fakhri, “New adaptation method based on cross layer and TCP over protocols to improve QoS in mobile ad hoc network”, International Journal of Electrical & Computer Engineering, 2021, vol. 11, no. 3, pp. 2088-8708.

M. N. Girish and A. Pande, “Improving TCP performance in wireless networks”, 2020.

S. P. Kuppusamy and M. Subramaniam, “Deep learning-based TCP congestion control algorithm for disaster 5G environment”, 2023.

M. R. Kanagarathinam, S. Singh, I. Sandeep, H. Kim, M. K. Maheshwari, J. Hwang, A. Roy and N. Saxena, “NexGen D-TCP: Next generation dynamic TCP congestion control algorithm”, IEEE Access, 2020, vol. 8, pp. 164482-164496.

E. Sy, T. Mueller, C. Burkert, H. Federrath and M. Fischer, “Enhanced performance and privacy for TLS over TCP fast open”, arXiv preprint arXiv:1905.03518. 2019.

T. Chowdhury and M. J. Alam, “Performance evaluation of TCP Vegas over TCP Reno and TCP NewReno over TCP Reno”, JOIV: International Journal on Informatics Visualization, 2019, vol. 3, no. 3, pp. 275-282.

A. Hernawan, “Comparative Performance Testing of the Impact of ACK Loss in TCP Tahoe, TCP Reno, and TCP New Reno on the ns-2 Simulator”, JOURNAL OF INFORMATICS AND TELECOMMUNICATION ENGINEERING, 2023, vol. 7, no. 1, pp. 91-101.

V. U. Devi and A. Thilaka, “A Comprehensive Analysis and Comparison of TCP Reno and TCP New Reno”.