A systematic performance evaluation method for a discrete-time tandem network

A simple and systematic method mainly composed of a decomposition technique and a parameter fitting algorithm is proposed in this paper to get end-to-end performance measures for a discrete-time tandem network with inputs of two-state Markov-modulated Bernoulli processes (MMBPs). To validate the proposed method, several experiments are conducted. We show that the proposed method has good accuracy in performance evaluation for such a tandem network in terms of delay, delay variance, and packet loss probability.