Traffic profiles for audio and data collaborative work systems

A predictive model of the behavior of multimedia communications systems on digital broadband networks is needed because, it is through knowledge of the traffic characteristics of the arrival process, that proper network resource dimensioning can be performed to avoid congestion and maintain the network quality of service. This thesis analyses the traffic from multimedia systems that are expected to become common in the next decade. The traffic is used to develop models to characterise the statistical properties of such multimedia communications systems.