12

Markovian Arrival Processes

12.1 Introduction

Modern telecommunication networks are designed to support multimedia traffic including voice, data, and video. One major feature of these networks is that the input traffic is usually highly bursty. Also, because these networks operate in a packet switching mode, there is usually a strong correlation between packet arrivals. Thus, for these networks, the traditional Poisson traffic model cannot be applied because the presence of correlation between traffic arrivals violates the independence assumption associated with the Poisson process. For example, in an asynchronous transfer mode (ATM) network, different types of traffic from different sources arrive at an ATM switch that statistically multiplexes ...

Get Markov Processes for Stochastic Modeling, 2nd Edition now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.