A nonstationary offered-load model for packet networks

Motivated by the desire to appropriately account for complex features of ne twork traffic revealed in traffic measurements, such as heavy-tail probabil ity distributions, long-range dependence, self similarity and nonstationari ty, we propose a nonstationary offered-load model. Connections of multiple types arrive according to independent nonhomogeneous Poisson processes, and general bandwidth stochastic processes (not necessarily Markovian) describ e the individual user bandwidth requirements at multiple links of a communi cation network during their connections. We obtain expressions for the mome nt generating function, mean and variance of the total required bandwidth o f all customers on each link at any designated time. We justify Gaussian ap proximations by establishing a central limit theorem for the offered-load p rocess. We also obtain a Gaussian approximation for the time-dependent buff er-content distribution in an infinite-capacity buffer with constant proces sing rate. The offered-load model can be used for predicting future bandwid th requirements; we then advocate exploiting information about the history of connections in progress.