Dimensioning of a transport network being robust to changes of the traffic pattern

Abstract

The planning and dimensioning of circuit-switched networks is investigated where the expected traffic pattern is uncertain. To model the vagueness, the static traffic matrix is replaced by a matrix with random variables for the predicted traffic. With the choice of the distribution for these variables, a scenario of different realizations may be characterized. A measure for the accuracy of the prediction is introduced. An algorithm for the dimensioning of the necessary network resources for realizations of this random process is presented and discussed. A case study for a pan-European network is performed.