Research On AOS Congestion Control Based On Self-similar Traffic Prediction

With the rapid increase of spatial network traffic,network architecture becomes more and more complex and network congestion becomes more and more serious.The occurrence of congestion will lead to the loss of network traffic,poor real-time performance and other problems,and even lead to network paralysis,which adds new complexity to the improvement of network service quality and reasonable allocation of network resources.At the same time,a large number of studies have shown that network traffic presents self-similar characteristics,which provides a new idea for the study of network congestion control.The predictability of self-similar traffic can be introduced into the study of network congestion control,so as to obtain more effective control strategies.This thesis mainly studies the AOS(Advanced Orbiting System)congestion control technology based on self-similar traffic prediction to achieve effective control of network congestion.The main contents of this paper are as follows:Firstly,congestion control algorithm of network traffic,self-similarity characteristics of traffic and AOS multiplexing model are studied,and a new congestion control algorithm of AOS under self-similarity traffic is proposed.This algorithm adds queue management module to AOS multiplexing model,and a new improved queue management algorithm is designed in this module.By detecting the average queue length of traffic,the congestion of the system can be controlled by corresponding discarding strategy.The simulation results show that the algorithm has a significant effect on the improvement of system performance.Secondly,the prediction method of self-similar traffic is studied.In this paper,the statistical prediction model is adopted,which uses the statistical method to deal with traffic correlation.The calculation process is relatively simple,and the model is easier to implement.On this basis,the Linear Minimum Mean Square Error(LMMSE)prediction algorithm is studied,the existing problems in the algorithm are corrected,and the correctness of the modified algorithm is verified by simulationFinally,AOS congestion control algorithm based on self-similar traffic prediction is studied.By introducing the predicted traffic into AOS multiplexing model and combining the modified LMMSE statistical prediction algorithm with queue management algorithm,an AOS congestion control algorithm based on self-similar traffic prediction is designed.Theoretical analysis and simulation results show that the algorithm has outstanding performance in congestion control,especially in improving the average delay and maximum delay of the system,and greatly reducing the system residual.This algorithm can provide theoretical support for engineering practice.