TCP with Machine Learning - Advances and Opportunities
Hardik K. Molia · International Journal of Advanced Trends in Computer Science and Engineering · 2019
TCP-Transmission Control Protocol is a connection oriented and reliable transport layer protocol to ensure process to process communication.Improving TCP's performance for wireless networks is always been a challenging task for the researchers.The primary reason of the TCP's performance degradation in wireless networks is its inability to differentiate losses to act accordingly.TCP suffers from inaccurate bandwidth estimation and fairness issue in wireless networks too.Over the years, researchers have proposed many TCP variants to enable TCP to maintain satisfactory performance in wireless environments.In the last decade, ML-Machine Learning techniques are started being applied in every possible area due to their numerous advantages.Many researchers have tried to apply various ML techniques to improve TCP's performance for wireless and wired networks.This paper discusses the fundamentals of ML techniques and their applicability with TCP.Some of the recent ML based TCP variants are discussed.The paper concludes with some future directions for researchers.