基于ARED的主动队列管理改进算法

随机早期检测（Random Early Detection,RED）是IETF推荐部署的主动队列管理（Active Queue Management,AQM）算法。 RED存在参数难以配置、无法适应动态网络环境的缺点。 ARED（ Adaptive RED）是RED的自适应版本,通过平均队列长度来动态调整最大丢弃概率,从而达到稳定平均队列长度的目的,但是存在瞬时队列长度振荡的问题。文中研究了拥塞控制中的主动队列管理,对ARED算法进行了改进,优化丢弃概率计算函数,提出TTS-ARED算法,实现在动态网络环境下队列长度的稳定以及丢包率降低。 NS2的仿真结果表明,TTS-ARED算法显著地降低了丢包率,队列长度稳定性比ARED算法更优越。

Improved Active Queue Management Algorithm Based on ARED

Random Early Detection （ RED） is the Active Queue Management （ AQM） algorithm recommended by IETF. RED is difficult to configure its parameters,and can＆#39;t adapt to dynamic network. ARED is an improved adaptive RED,by measuring average queue size to dynamically adjust maximum drop rate,achieving stable average queue size. But it may cause instantaneous queue size oscillation. It re-searched the active queue management in congestion control,and some improvements are made based on ARED algorithm. It optimized the calculation of dropping probability and a new algorithm TTS-ARED is proposed,low drop rate and the stable queue size in dynamic network are realized. The simulation results indicate that the drop rate can be decreased significantly by using TTS-ARED,and in the area of stabilizing queue size,TTS-ARED algorithm is superior to RED algorithm.