Research On Congestion Control Mechanisms In High Latency Hybrid Networks

In recent years,with the deep integration of the fourth industrial revolution and various Internet communication technologies,wireless networks with high bandwidth and high latency characteristics are being integrated into high-speed wired networks,resulting in a high-latency hybrid network with high heterogeneity and high transmission delay characteristics.Due to the uneven allocation of network resources and congestion caused by the large amount of data from the proliferation of users,which can cause significant damage to network transmission performance,active control of network congestion is particularly important.However,congestion control mechanisms designed for traditional wired networks do not perform well in high-latency hybrid networks,and the large deployment of memory in network devices and TCP congestion control mechanisms to avoid packet loss can cause serious bufferbloat problems.Active Queue Management(AQM),as a queue management mechanism to assist TCP congestion control,has been widely studied because of its ability to suppress congestion.Therefore,this thesis introduces and improves the AQM congestion control algorithm to address the bufferbloat problem and high buffer queue tail delay in high latency hybrid networks,and designs an AQM feedback controller in conjunction with classical control theory to further reduce network congestion and improve network transmission performance to ensure efficient and smooth operation of high latency hybrid networks.The thesis focuses on the above analysis and contains the following main aspects.Firstly,the network congestion control strategy is studied,and the causes of bufferbloat problem and high buffer queue tail delay problem in high-latency hybrid networks are analyzed in depth from two aspects: network structure characteristics and congestion control algorithm,and the shortcomings exposed by traditional TCP congestion control algorithm are analyzed in detail according to the design principle of network congestion control strategy to give high latency In addition,we also analyze the impact of the shortcomings of traditional TCP congestion control algorithms on the congestion control mechanism in high-latency hybrid networks,and summarize the research proposal based on this thesis.Secondly,by analyzing the Co Del algorithm in a high-latency hybrid network environment,the Kalman filter-based Co Del algorithm(Co Del-K)is proposed to mask these abnormal signals by addressing the problems of noise interference and burst data streams in the network that affect the network transmission performance.The optimization algorithm is also mathematically modeled and differential equations are modeled to understand the network dynamic characteristics of the algorithm.By using discrete-time harness theory to theoretically analyze the network system state,it is proved that the Co Del-K algorithm can ensure the stability of the network system while improving the portability of the algorithm.The effectiveness of the algorithm is verified through NS-3 simulation experiments to ensure efficient network transmission.Finally,an active queue management algorithm based on classical control theory is proposed for a complex and variable time lag system like TCP/AQM in a high latency hybrid network environment.Based on the small-signal linearization of the nonlinear TCP/AQM fluid model,a Co Del-K controller based on AQM feedback control is designed,and the good steady-state performance of the TCP/AQM network feedback system is demonstrated by the frequency domain analysis method.MATLAB simulation results show that the algorithm has good control capability and anti-interference capability in the high latency hybrid network environment.And the NS-3 simulation experiment verifies that the algorithm has excellent performance in the dynamic network environment.