可用带宽是重要的网络资源,但实际测量却存在许多困难。基于Cruz流量模型和单向时延OWD(one—way delay)趋势分析,提出一种端到端的可用带宽主动测量算法PathPCQ(Path Parameter based on Cruz Queue)。PathPCQ通过主动检测OWD连续上升的概率,推断探测速率与端到端可用带宽之间的关系。当探测速率大于可用带宽时,其概率大于阈值;否则小于阈值。由此调整探测速率,使之逼近可用带宽。ns2仿真结果显示,PathPCQ算法测量精度和效率都优于现有的带宽测量工具Pathload,为网络资源管理、监控和预测提供了有益的手段。
Available bandwidth is one of the important network resources but it is difficult to measure in practice. PathPCQ (Path Parameter based on Cruz Queue), an active algorithm for end-to-end available bandwidth estimation is proposed. In PathPCQ, the probability of the successive rising one-way delay is statistic according to Cruz traffic model and one-way delay trends to infer the relationship between the probing rate and the end-to-end available bandwidth. If the probability is larger than the threshold, the probing rate is larger than the present available bandwidth; otherwise, the probing rate is less than the present available bandwidth. By adjusting the probing rate, the available bandwidth can be estimated. The ns2 simulation results show that PathPCQ is more accurate with less convergence time than the present estimation tool Pathload. PathPCQ provides a beneficial method for the network design, management, control, and prediction.