摘要
提出了一种基于速率的主动队列管理机制。主要思想是根据每流过去的输入速率和输出速率预测下一个采样间隔内的输入速率和输出速率 ,用上一个采样间隔末的每流队列长度、预计输入、预计输出和期望的队长来决定数据包的丢弃概率 ,达到稳定队列长度和实现公平性的目的。仿真结果证明在多种网络条件下 。
This paper presented a speed-based active queue management mechanism(SBAQM). In this mechanism, the incoming and outgoing speed of each flow in the next sample intervals can be predicted according to the past data, and the drop probability of some internal can be calculated by the queue length of each flow at the end of last internal, predicted incoming and outgoing data and expected queue length. Through ns-2 simulation, SBAQM was compared with FRED and SRED. The simulation results show that SBARQ can achive stable queue length and per-flow fairness.
出处
《计算机应用》
CSCD
北大核心
2004年第8期82-85,共4页
journal of Computer Applications
基金
国家自然科学基金项目 (90 2 0 4 0 0 8)
