分组交换中分组流的概率特性与输出排队

The Probability Characteristics of the Traffic Stream and the Output Queues in a Packet Switch

本文分析快速分组交换中分组流的概率特性与输出排队。在输入分组流为复合泊松流的假设之下，论证了复合泊松流在分组交换过程中的叠加性、分解性、泊松性、马尔柯夫性、输出与输入的不变性等。然后，将输出分组流转换为连续时间的马尔柯夫链，分析了输出分组流的概率特性，并得到了输出排队长分布和充满缓冲器的概率。

This paper disscusses the probability characteristics of the traffic stream and the output queues in a fast packet switch system.Under the assumption that complex Poissin Flow is regarded as input traffic stream, the property of superposition,disassembly,Poisson,Markov and constant of output and input in a fast packet switch are proved.Then output traffic stream is converted into continuous time Markov Chain,the probability Characteristics of output traffic stream are analysed and the output queues length distribution and full buffer probability are obtained.