An empirical comparison of diffusion approximations and simulation in ATM networks

We consider an ATM switch model, in which each of N sources is a Markov modulated rate process. We look at some approximations that have been proposed for p, the probability that in steady-state conditions, the buffer content exceeds x. These approximations are compared with the simulation results for p and the cell-loss ratio.

[1]  D. Iglehart.  Limit Theorems for the Multi-urn Ehrenfest Model , 1968 .

[2]  Y. Champoux,et al.  Importance sampling for large atm-type queueing networks , 1996, Proceedings Winter Simulation Conference.

[3]  H. Kobayashi,et al.  Diffusion process approximations of a statistical multiplexer with Markov modulated bursty traffic sources , 1994, 1994 IEEE GLOBECOM. Communications: The Global Bridge.

[4]  Vidyadhar G. Kulkarni,et al.  Fluid Model Driven by an Ornstein-Uhlenbeck Process , 1994, Probability in the Engineering and Informational Sciences.

[5]  S. Schach.  Weak Convergence Results for a Class of Multivariate Markov Processes , 1971 .

[6]  Krzysztof Debicki,et al.  A Gaussian fluid model , 1995, Queueing Syst. Theory Appl..