Intelligent Congestion Avoidance in Differentiated Service Networks

Active Queue management (AQM) takes a trade-off between link utilization and delay experienced by data packets. From the viewpoint of control theory, it is rational to regard AQM as a typical regulation system. Although PI controller for AQM outperforms RED algorithm, the mismatches in simplified TCP flow model inevitably degrades the performance of controller designed with classic control theory. The Differentiated Service (Diff-Serv) architectures are proposed to deliver Quality of Service (QoS) in TCP/IP networks. The aim of this paper is to design an active queue management system to secure high utilization, bounded delay and loss, while the network complies with the demands each traffic class sets. To this end, predictive control strategy is used to design the congestion controller. This control strategy is suitable for plants with time delay, so the effects of round trip time delay can be reduced suing predictive control algorithm in comparison with the other exciting control algorithms. Simulation results of the proposed control action for the system with and without round trip time delay, demonstrate the effectiveness of the controller in providing queue management system.