Abstract:

From the angle whether maintaining flow state information, this paper presents an improved Active Queue Management(AQM)——Single Flow-AQM(SF-AQM). SF-AQM only maintenances data flow state information which has high transmission rate in order to reduce the router overhead identifies the non-adaptive flow by comparing packet the arrival intervals of different flow, improves the fairness of the algorithm, and controls the queue length under the target value to ensure the stability of the algorithm. Simulation results show SF-AQM algorithm has good performance in fairness and stability, and it is an effective algorithm in high performance communication networks congestion control.

Key words: congestion control, Active Queue Management(AQM) algorithm, fairness, queue length, bandwidth utilization rate

摘要：

从是否维护数据流状态信息的角度出发，提出一种改进的主动队列管理算法——SF-AQM。SF-AQM算法只维护发送速率大的数据流状态信息以降低路由器的开销，通过比较不同数据流的包到达时间间隔衡量流到达速率，识别出非适应性数据流，提高算法公平性，并使队列长度控制在目标值附近，保证算法稳定性。仿真结果表明，SF-AQM算法具有较好的公平性和稳定性，且对抑制网络拥塞有明显效果。

关键词: 拥塞控制, 主动队列管理算法, 公平性, 队列长度, 带宽利用率

CLC Number: 

• TP393.02