Interval characteristics of group Poisson models of telecommunication systems traffic

Lack of efficacy of applying self-similar processes models to the analysis of queues in telecommunication systems is shown. The evolution of flow models controlled by the Markov chain is considered. The features of applying Markov flows as traffic models of telecommunication systems are considered. Models of single-channel queuing systems with input flows with arbitrary correlation are presented. Generalizations of the Khinchin-Pollachek formula are given for these systems. The prospects of using interval methods of queue analysis developed by the author in queuing systems with correlated input flows are shown. It is proposed to use the group nonordinary Poisson flow as the model of telecommunication traffic. Interval characteristics of these flows are considered and the prospects of their application are shown. The questions of multiplexing such flows during processing in queuing systems are considered. It is shown that when summing several group Poisson flows, the resulting flow is also a group Poisson flow. The conclusions made are confirmed by the results of simulation. By examples, the adequacy of characteristics of such models to the characteristics of real flows of video traffic is shown.