Modelo para a distribuição de probabilidade do comprimento dos pacotes em redes de computadores.

Abstract

Thetraditionalmodelsoftelephonetrafficarenotsuitableforpackettrafficincomputer networks. Therefore, research to find new models focused on the study of several features associated withduration offlow, timebetween packet arrivals, packet lengthsand theirstatistical properties. It was a seach for more suitable equations for its modeling. This thesis presents a model that leads to an equation for the probability density function, and its cumulative distribution function, of the packet length for computer network with bimodal traffic. This type of trafficis important because it appears in the literaturein several actual measurements of packet length. The model can be used for synthetic traffic generation, comparison, simulation and estimation of traffic on computer networks. The proposedmodel is validated by comparisons with actual data traffic measured by the author, experimental results of the literature and comparisons with other probability distributions.