TRAFFIC MODELING AND STATISTICAL ANALYSIS IN NETWORK TOMOGRAPHY

ABSTRACT. The estimation of traffic matrices in a communications network on the basis of multi-period traffic measurements on network links is an important problem, and several solutions have been proposed when the traffic does not show dependence over time. However, extensive measurements cam-paigns conducted on IP networks have shown that the traffic actually exhibits long range dependence. Here a method is proposed for the estimation of traffic matrices in the case of long range dependence, and its properties are studied. 1 INTRODUCTORY ASPECTS A telecommunication network is ideally represented as a set of nodes (each corresponding to a transmission tool) connected by arcs (physical links). The set of nodes and the corresponding links define the topology of the network. Time is discretized in time slots. At each time slot, each node (S: source node) sends information (bytes) to every each other node (D: destination node). If the link between nodes O and D exists, then information from S to D flows through such link. Otherwise, information is routed through a path of consecutive links connecting S and D. The vector containing the expected amount of traffic per time slot from each source nod