一种基于负载的公平性主动队列管理算法

针对随机早期检测（RED,Random Early Detection）算法存在的公平性问题,提出了一种基于负载的公平性主动队列管理算法（LFED）。该算法通过引入网络负载和队列这两个概念,有效地判断当前网络拥塞程度,同时使用改进的丢包率公式,并且借鉴CHOKe算法的惩罚机制对非响应流进行有效惩罚,以保证不同数据流之间的公平。仿真实验表明,与RED和CHOKe相比,LFED可以得到稳定的瞬时队列长度和减少分组丢弃率,具有较好的公平性和稳定性。

Load-based Fair Active Queue Management Algorithm

For the fairness problem in Random Early Detection（RED） algorithm,this paper proposes a load-based fair AQM algorithm—LFED.This algorithm,through introduction of network load and queue,could effectvely identify the network status.LFED,with an improved packet loss rate formula and by reference of CHOKe punishment mechanism,imposes effective punishment upon non-responsive flows,thus to ensure the fairness of different flow.The simulation results show that,compared with RED and CHOKe.LFED could brings more stable queue length and decrease packet loss rate effectively,and present better fairness and stable performance.