Implementation Of AQM Algorithms By FPGA Based On DSPACE Network Model

Nowadays, problem can not be ignored with the development of network. Thenumber of users increases cause a large network burden, directly leads to networkcongestion problem. In order to solve this problem, research on AQM has becomemore and more deep. Many algorithms of traditional control field have beenintroduced to network congestion, also very effective. But complex algorithms lead tomore router hardware resource is occupied, performance of the network is poor.FPGA has the advantages of computing speed fast, on-chip resource rich, so theresearch on implementation of AQM algorithms by FPGA is very meaningful.This paper mainly research of implementation of AQM algorithms by FPGA.The fuzzy neuron AQM algorithm is designed combined with the theory of fuzzycontrol and neuron control, then implement AQM algorithms by FPGA with idea ofSoPC scheme based on Nios II softcore processor. The network model is designed toverify the effect of AQM algorithm implementation of FPGA. The innovativecontents of this article are as follows:1. Research of network congestion mechanism and a variety of AQM algorithm,designed a fuzzy neuron algorithm based on neuron control and fuzzy control.2. Research of FPGA development formula, choose the SoPC developmentprogram based on NiosⅡ, implementation AQM algorithm by FPGA with thisprogram. Describes a method of design of hardware engineering and softwareengineering.3. In order to test and verify the effect of of the controller (the AQM algorithmimplemented by FPGA), describe the TCP network by formula in Matlab/Simulink,and download it into dSPACE as the target.4. Establish communication module of FPGA and dSPACE, build anexperimental platform, analyze the experimental results by simulation.The experimental results demonstrate that it is feasible to implement AQMalgorithm by FPGA, base on the SoPC development program. The controller cancontrol network model work stable through investigating the queue length trend ofnetwork model, good control effect has been reflected, and ability of againstinterference is great too. The AQM algorithm implemented by FPGA can solve thenetwork congestion problems better, and this scheme has the characteristics ofminiaturization, high reliability and low cost.