基于Cauchy分布的抖动模型

Jitter model based on Cauchy distribution

为了准确地刻画IP网络中的抖动,提出了一个基于Cauchy分布的抖动模型。同时根据该抖动模型,说明了随着路径中的节点的增多,网络抖动的离差与节点数目之间呈密率关系。可以利用这一结论,通过节点的数目和流量参数的大小来预测应用的QoS,并可以进行估算,为了保证应用特定的QoS,沿网络传输所能允许的最大节点数。仿真实验结果表明,Cau-chy模型与实际测量值之间可以得到很好的吻合。

To accurately depict the jitter of IP network, a jitter model based on Cauchy distribution is proposed. Furthermore, according to the proposed jitter model, the relation of the dispersion of network jitter and accumulated network nodes shows power law, as the nodes of the path is accumulated. This conclusion is used to predict QoS according to node number and flow parameter. And it is used to calculate max node number along the transmission path, in order to promise special QoS of applications. Simulation results show Cauchy model and the practical measured values can match very well.