Real Time Data Traffic Analysis Using Poisson Process in Next Generation Network

Most of the applications in the Next Generation Network are real time multimedia applications. Quality of Service (QoS) is a major constrain in those type of applications. In order to provide good quality video, audio etc. in the real time applications, the traffic in the network should be well modelled. So that the two metrics of QoS that are Call Blocking and Call Dropping Probability can be reduced. Mostly multimedia traffic in the network is bursty in nature. With the identification of both Short Range (SRD) and Long Range Dependence (LRD) in bursty traffic, the modelling in high speed network has become a challenging one. In this paper, the classical memory less Poisson model and the self- similarity model are analyzed and showed why Poisson process is inappropriate in modelling the high speed network traffic.