Traffic models with a rate varying according to a Gaussian distribution are commonly used to evaluate statistical multiplexing in telecommunication systems. The Ornstein-Uhlenbeck process (OUP) is a Markovian and Gaussian process closely related to a first order autoregressive process (AR(1)). The superposition of voice or other homogeneous traffic sources asymptotically approaches an OUP in continuous time or an AR(1) in discrete time models, respectively. Performance results for a multiplexer with OUP input (OUP/D/1) are obtained based on functions of a single system parameter, We use fluid flow analysis and an approach for discrete time semi-Markovian systems to analyze the OUP/D/1 queueing system. A comparison of waiting time quantiles for input from a limited number of voice sources with the OUP/D/1 results indicates rather small deviations already for a moderate number of sources.

Det tas fram en approximation för väntetiden som överskrids av bara 1 % av alla röstpaket som måste köas i en flaskhals. Approximationens kvalitet visas genom jämförelse med flödesmodellberäkningar.