Connection admission control for video traffic using modified equivalent capacity

For realization of reliable ATM networks, video traffic modeling and traffic control are required to prevent network congestion. Connection admission control (CAC) is one of important technologies of traffic control. A method for calculating the equivalent capacity of an aggregated traffic source is proposed. We assume individual traffic sources as autoregressive (AR) models and compute the first, second, and third moments of the total bit rate of the aggregated source, as a function of the moments of individual traffic sources. We then approximate the aggregated traffic by the 2-state Markov modulated Poisson process (MMPP) and obtain the 2-state MMPP parameters from the computed moments of the aggregated traffic. Next we calculate the equivalent capacity of the 2-state MMPP by the proposed modified equivalent capacity. Computer simulation shows that the proposed method for calculating the equivalent capacity has a higher statistical multiplexing gain compared with conventional methods.