Simplifying Blocking Probability Estimations for Heterogeneous Voice Sources in VoIP Environments

Abstract

A fast estimation of blocking probability for heterogeneous voice source in VoIP environments is proposed in this paper. In VoIP environments, traffics are generated from various voice sources and the transient system states due to the input traffics are commonly described by MMPP (Markov-Modulated Poisson Process). MMPP is a generalization of Poisson process and is commonly used in modeling the input process of communication networks. However, as we know, the solution of MMPP model on the estimation of blocking probability is very involved when the total system state number N become large. In this paper, a generic reduction method is proposed for blocking probability estimation. In our study, a large total number of system states for the original MMPP model is equivalently downsized to an approximating Markov chain model with less total number of system states. This will benefit on easy estimation of blocking probability. Our contribution is focused on system simplification and the blocking probability estimation is asymptotically closed to the original system. The numerical results shows our method is satisfied and this discipline can be widely applied to other high-speed networks for model simplifications.