RED队列稳态误差分析

Steady State Error Analysis of RED Queue

动队列管理在保证较高吞吐量的同时 ,通过在交换节点上主动丢弃数据包来控制队列长度 ,从而实现对端到端的延时和抖动的控制 RED算法是目前应用最为广泛的主动队列管理 (AQM )算法 RED算法以平均队列长度作为衡量网络拥塞的指标 ,其参数设置对算法性能有较大影响 利用现代控制工程理论 ,将RED算法看做一种单位反馈控制系统 ,并将期望队列长度作为系统输入 ,将瞬时队列长度作为输出 ,对该系统的稳态误差进行了分析 实验结果表明在稳定状态下 ,RED队列的波动受分组丢弃概率函数的斜率影响 在稳定条件边界附近 。

Active queue management (AQM) is used to drop the packets actively on the intermediate nodes, so as to control the queue length, in order to control end-to-end delay and jitter while keeping high throughput. As one of the AQM mechanisms, RED is the most widely implemented on the routers. RED takes the average queue length as the main measure of congestion and its parameters have great effect on the algorithm performance. RED is seen as a unit feedback control system. Taking expected queue length as input and instantaneous queue length as output, steady state error of this system is analyzed. The simulation result shows that the slope of the packet's discarding function influences the queue fluctuating. When closing to the border of the stable conditions, the steady state error of the system will increase rapidly.