A Proportional-integrative(PI) Controller Based On Multi-rate Sampling For Congestion Controlling

To analyse internet congestion control have many difficulties such as: 1) internet is a very complicated system with complex network architecture, huge scale. The application type is evolving unceasingly. the number of Network users changes often and all kinds of random fault may happen sometimes. so it causes the Internet modelling to be extremely difficult. 2)It is impossible on Internet to use the common control, but to use the disperser feedback control mechanism. 3)When feedback signal transmission inevitably exist transmission extends, Moreover along with the path and the environment differently, extends will also unceasingly changes. 4)As a result of application and environment difference,we needs to use the different congestion control strategy. This causes to analyze each kind of control strategy the mutual influence as well as they changes to the entire network stable influence difficultly. For all this, Recently in several years, the massive work on analysising exist congestion control in the use the stable state and the dynamic performance as well as design the new congestion control algorithm aspect have done and has obtained the good beginning.With regard to the flow regulation of the best-effort traffic,the controllable traffic in high speed computer communication networks,the present paper proposes a novel control theoretic approach that designs a proportional-integrative(PI) controller based on multi-rate sampling for congestion controlling.Based on the traffic model of a single node and on system stability criterion,it is shown that this PI controller can regulate the source rate on the basis of the knowledge of buffer occupancy of the destination node in such a manner that the congestion-controlled network is asymptotically stable without oscillation in terms of the buffer occupancy of the destionation node;and the steady value of queue length is consistent with the specified threshold value. We use a colorful noise to describe traffic flow and regard it as the uncontrolled input(equaled to noise).Then we use the multi-rate sampling PI controller to eliminate the error caused by noise, and make the length of queue stabilize at the threshold.The availability of the strategies and stability of systems are proved by doing some analysis and simulation.