基于速率自适应预测的ATM网络ABR业务流量控制

ATM Network's ABR Service Flow Control Based on Adaptive Rate Prediction

本文尝试将自适应预测方法引入ABR流控算法中。ATM网络的ABR业务带宽模型是一非平稳随机过程。高速网络中可用带宽的反馈信息由于线路的传播时延造成了流量控制的滞后。本文充分考虑了ABR业务的这两个特点，对ATMForum的TM4．0规定进行了一定的修改，采用离散化的ATM网传输模型，提出了一种基于速率自适应预测的ABR业务流控方法，较好地补偿了时延带来的控制的滞后效应，且具有参数易于选择的优点。通过仿真，论证了该方案的有效性。

This paper brings the adaptive algorithms into the ABR flow control mechanism. ABR service in ATMnetwork is a non-stationary stochastic process. Because of the propagation time, the available bandwidth information in highspeed network leads to the delayed effect of the flow control. To solve these two problems, we employ the discretetransndssion model of ATM network to present a ABR flow control mechanism based on rate adaptive prediction. By meansof this mechanism, the delayed control raised from delayed feedback information can be compensated and the operatingparameters can be easily chosen. The effectiveness of this mechanism has been verified by computer simulation.