Atm Network Abr Congestion Control, Control Theory Algorithm Analysis And Research

Flow/congestion control is the critical problem of ATM traffic management, and congestion control of ABR service is its research issue. The main drawback of classic congestion control algorithms for ABR service is that they are based on heuristics that do not have a formal theory to back them and/or depend on a number of measurements that are either difficult to perform, or have a significant amount of noise in their estimates. Filtering out the noise will lead to sluggish response, with higher cell loss and lower link utilization. Further, these proposals do not provide any formal design methodology to ensure the stability of the control loop. Then, a number of algorithms based on control-theoretic principles have appeared in recent years.The intention of this paper is to reach a conclusion about algorithms of the ABR congestion control theory. It also presents both the advantages and disadvantages in these algorithms. In the second part of this paper, it makes a detailed analysis among fluid-model-based, smith-principle-based and two-PD-controllers-based by comparing to the method and application of control theory. Then it gets the laws to keep the control loop stable. In order to analyze the dynamic and transient characters, three congestion control algorithms mentioned above have been carried out in the NIST ATM network simulator and the successful outcomes are brought out. Some research topics in this field are also presented in the final part of this paper.