The statistics and analysis of packet traffic distribution characters in access network based on NetMagic platform

Packet traffic has random characters and relates to different layers. This paper aims to analyze the statistics characters of real packets on service node interface (SNI) of access network, using complex Poisson distributed probability model to describe the characters of the number of received packets, and using tri-modal distributed probability model to describe the characters of the packet length. K-S fit tests are used to evaluate these models and achieve high accuracy results. Based on these models, a probability model of packet traffic is proposed. The methods can be extended to research higher layer network traffic characters.