Two-state MMPP modeling of ATM superposed traffic streams based on the characterization of correlated interarrival times

We propose a new two-state Markov modulated Poisson process (MMPP) modeling method based on the characterization of interarrival times in heterogeneous traffic environments. We identify the two state MMPP parameters by matching the distribution and autocovariance functions of interarrival times. Our proposed modeling method can efficiently approximate the ATM superposed traffic and it yields upper estimates of cell loss ratio requirements. The proposed two-state MMPP modeling can be utilized in the traffic characterization of ATM traffic streams and connection admission control.