Assoc. Prof. Seferin Mirtchev
15 July 2010
IP networks have become a dominant factor in bringing information to users on a worldwide basis and it is very important to provide QoS. There are numerous traffic models proposed for understanding and analyzing the traffic characteristics of networks. But there is no single traffic model that can efficiently capture the IP traffic characteristics of all types of networks. This presentation is one step towards this goal. We have offered different distribution as: Generalized Poisson Process, Generalized Erlang, Pareto and Polya. With these distributions we can define different behavior of the queuing systems by two parameters: mean value and variance. We have discussed and presented numerical results for some continues time queuing systems as: M(g)/M/1/k/S, M/M(g)/1/k/S, Pareto/D/1/k and Polya/D/1/k and also for some discrete time traffic models as: Geo/MM/1/k, Pareto/MM/1/k, Geo/D/1/k and Geo/Polya/1. We have shown that the presented distribution and queuing systems are more proper for IP traffic analyses.