Statistical delay analysis on an ATM switch with self-similar input traffic

A self-similar stochastic process was used to characterize the arrival traffic for an asynchronous transfer mode (ATM) switch. With this approach, the statistical delay bound on the worst case cell delay can be determined. The statistical delay analysis provides the methods to determine the statistical delay bound for the network traffic which is more suitable for the connection admission control for an ATM network with different output port controllers. LAN traffic experiments show that the statistical bound is tighter than the delay bound derived from previous methods and stays close to the actual maximum cell delay in an ATM network.