Efficient estimation of cell blocking probability for ATM systems

Two contributions that provide the foundation for the application of IS (importance sampling) to the ATM (asynchronous transfer mode) queueing process are presented. First, a combination of IS and Monte Carlo simulation is used to estimate the cell loss probability. Second, it is recognized that the conditional arrival process must be biased directly to obtain significant computational improvements. The results show that if applied properly, IS can be useful in ATM system analysis; typically, the computational burden can be reduced by more than three orders of magnitude. The foundation given for applying IS to ATM systems can be used to study congestion control as well as future networks of ATM queues