Research And Improvement Of MPTCP Transmission Technology Based On Heterogeneous Mobile Internet Network

With the rapid development of various access technologies,mobile communications,satellite communications,etc.,heterogeneous networks composed of different computers,mobile terminals,network devices,and systems that integrate multiple networks have become the inevitable direction of future network development.As an extension of the TCP protocol,Multipath Transmission Control Protocol(MPTCP)can utilize multiple ports of terminals accessing heterogeneous networks to aggregate multiple transmission paths for concurrent transmission,thereby improving transmission throughput and stability.However,in heterogeneous networks,network transmission based on the MPTCP protocol faces technical issues such as path management,congestion control,and buffer configuration.Simply solving one of these issues often cannot achieve good results,which collectively affect the performance of MPTCP transmission.Based on this background,this thesis studies the MPTCP transmission technology in heterogeneous networks and proposes comprehensive improvement methods.This thesis provides an overview of the research status of MPTCP protocol and its key technologies both domestically and internationally,and analyzes different path management algorithms,congestion control algorithms,and the relationship between buffer size and throughput through experimental testing.The experimental results indicate that the path management,congestion control,and buffer configuration of MPTCP interact and collectively affect the throughput of network transmission.In response to this experimental result,the paper proposes a multi parameter comprehensive optimization algorithm based on machine learning,which comprehensively optimizes path management,congestion control,and buffer configuration.The main content of this thesis as follows:(1)Through extensive experimental testing and analysis,key factors affecting transmission throughput have been identified,including path management,congestion control,and buffer configuration;(2)Propose a multi parameter comprehensive optimization algorithm based on machine learning,which consists of three modules: throughput prediction module,buffer prediction module,and adaptive buffer configuration module.The algorithm comprehensively improves and optimizes the three key factors that affect throughput;(3)Construct multiple different scenarios and cases on the NorNet test bed and Mininet test platform to conduct performance testing and analysis on the algorithm proposed in this article.Through experimental testing of multiple scenarios and cases,the results show that the algorithm proposed in the paper can not only select available paths and congestion control algorithms,but also accurately calculate the buffer size required for transmission scenarios,effectively saving a large amount of buffer resources while ensuring maximum transmission throughput.Compared with traditional formulas,the maximum savings in buffer resources can reach over 60%,At the same time,it can adaptively change the configuration of buffer size to adapt to changes in the network environment in the context of network fluctuations.