一种基于两级线性平滑的RED改进算法

An Improved RED Algorithm Based on Two-Level Linear Smoothing

通过对随机早期检测(RED)算法的分析,提出了一种基于两级线性平滑的改进算法——TLLS-RED。算法的基本思想是在最小阔值和最大阔值的中点增加一个中间阔值,使得当平均队列长度处于最小阔值和中间阔值之间时,丢包率增长较慢,处于中间阔值和最大阔值之间时,丢包率增长加快,有效地提高了拥塞控制能力。通过NS2仿真结果表明该算法明显提高了网络吞吐量、链路利用率和路由器缓冲区利用率等性能,增强了网络的鲁棒性。

Through analyzing the random early detection(RED)algorithm,an improved algorithm based on two-level linear smoothing is proposed.The basic idea of algorithm is to add an intermediate threshold in the midpoint of the minimum threshold and maximum threshold.When the average queue length is between the minimum threshold and intermediate threshold,the loss rate growth slowly,and the loss rate growth rapidly in the maximum threshold.It effectively improves the congestion control.NS2 simulation results show that the algorithm significantly improves the network throughput,link utilization and router buffer utilization and other properties,and also enhances the robustness of the network.