Effects of ON-OFF Variability in Two-State Pareto Traffic Models on Multimedia Application Transmission Performance

Formulae of multimedia transmission buffer overflow probabilities are developed for an ON-OFF two-state Pareto multimedia source. Then, these are used to study the effect of ON-OFF variability on the multimedia application transmission impact. For the same utilization level, the variability of OFF periods tends to worsen the overflow performance, while that of ON periods tends to ease the buffer overflow. The variability of OFF period has the total different effect son the multimedia application transmission performance from that of ON period. This paper has developed an analytic approach to predict the buffer overflow probabilities for an ON-OFF Pareto multimedia source. Though the arrival process is far from Poisson, the proposed embedded Markov chain approach provides a very accurate prediction.Furthermore, the results of the approach are used to explore the effects of variability in ON-OFF periods on multimedia application transmission performance.