Study On AOS Active Queue Management Technology Based On Self-similar Traffic

With the development of communication networks,satellite communications play an increasingly important role in social production and development.For example,emergency communications and disaster relief,network convergence and development of national defense and military affairs are all inseparable from satellite communication technologies.Advanced Orbiting System(AOS)protocol is the most widely used protocol in satellite communications networks.In this paper,AOS protocol active queue management algorithm,feedback mechanism and virtual channel scheduling algorithm are studied.First of all,this paper introduces the AOS protocol data unit format,multiplexing model and self-similar principle.AOS uses multiplexing mechanism to allow multi-user shared channel resources.In the frame generation algorithm,data packets from different sources are encapsulated into frames,and the channel is divided into multiple virtual channels according to the data type.Each virtual channel stores data with the same or similar characteristics,and then transform the data through a scheduling algorithm.Then this paper introduces several kinds of queue management algorithms such as random early detection algorithm and their advantages and disadvantages.Random early detection algorithm is applied in computer networks more queue management algorithm,but can not be directly applied to AOS.This paper presents a time-based active queue management algorithm for self-similar traffic.The algorithm uses the ON / OFF model to generate a self-similar traffic.The arrival of a source is regarded as a Poisson process,and the drop probability of the packet is calculated by counting the duration of sending data continuously by each arriving source.The longer the data is sent,the greater the discard probability is.The simulation results show that the algorithm can control the average cache queue length.Finally,this paper proposes an active queue management algorithm based on feedback mechanism.In the process of virtual channel scheduling,when the discard probability is calculated in the queue management algorithm by counting and recording the information served by each virtual channel,the calculation of the discarding probability is performed by combining the time length of the data sent by the source with the information.Feedback mechanism can improve the performance of the system,such as reducing channel delay,system residuals and frame loss rate.