Simulation Analysis And Research On Network Data Transmission Congestion Control Mechanism

At present,with the rapid development of information technology and the development and integration of network and industry,the scale of the network continues to expand,and more and more network applications not only require higher throughput,but also stricter requirements on delay.At the same time,a large number of multimedia services bring a large amount of network data traffic.When the network traffic is large,the buffer of the network forwarding node is easily filled,and then network congestion occurs,resulting in queuing delay and low throughput,reducing user experience and even causing the network to crash.The current network congestion control is an interactive collaboration between the TCP(Transmission Control Protocol)and the Active Queue Management algorithm,which plays an important role in alleviating network congestion and improving network service quality.This paper first investigates classic network congestion control algorithms,including TCP congestion control algorithms based on end nodes and queue management algorithms based on network intermediate nodes,and analyzes and introduces their principles.The TCP congestion control algorithm is mainly divided into loss-based,delay-based and hybrid mechanism types.According to the information fed back by the receiving end to perceive the network congestion state,the congestion control window size is adjusted to avoid network congestion caused by too fast packet transmission.Queue management algorithms are divided into active queue management algorithms and passive queue management algorithms.The active queue management algorithm actively drops the data packets in the intermediate node buffer before the congestion occurs,and feeds back the congestion signal to the source to reduce the data transmission rate,thereby reducing the possibility of network congestion.Then,based on the NS3(Network Simulator 3)simulation platform,this paper builds wired network and wireless Wi-Fi network scenarios,and discusses the performance differences achieved when the seven TCP congestion control algorithms are combined with the PIE(Proportional Integral controller Enhanced)algorithm.By designing scenarios with different congestion levels,the simulation research a combination of TCP/PIE algorithms that perform better in different network scenarios.At the same time,based on the lightweight active queue management algorithm PIE,aiming at the two main parameters it maintains,that is,the queue reference delay and the update cycle of the droped probability p,the impact of different parameter settings on the performance of the PIE algorithm is discussed.The simulation analysis is also carried out in the wired network and wireless Wi-Fi network scenarios.The results show that different parameter settings will affect the performance of the PIE algorithm,and the difference is more obvious in the wireless Wi-Fi network scenario.Finally,this paper proposes a fine-grained proportional integral queue management algorithm CPIE based on the PIE algorithm to address the current active queue management algorithm’s lack of fine-grained analysis of data streams and undifferentiated packet drop policies that lead to the inability to quickly alleviate network congestion in the case of network congestion.The proposed algorithm takes advantage of the Cuckoo hash table query speed and high load factor,and according to the application needs,it is modified as a Cuckoo counter and introduced into the queue management algorithm PIE as a traffic measurement module to store and provide efficient frequency queries for mapping key information of data streams,which enables fine-grained identification of data streams for differential packet drop policies in case of network congestion.Experimental results on the NS3 simulation platform show that the C-PIE algorithm achieves lower average queue length and packet round-trip delay than the RED,Co Del,and PIE algorithms,and maintains higher fairness and better stability in both wired and wireless Wi-Fi network scenarios,while maintaining similar throughput performance.