A scalable approach for steady state traffic modeling in high-speed backbone networks

This work proposes a high modular and scalable framework to effectively represent the average behavior of highspeed backbone networks, which usually deliver large volumes of highly aggregated traffic. The proposed framework combines together suitable TCP and network layers' models to effectively represent both the steady-state backbone network behavior and the performance level perceived by users. In Our approach is composed by three different analytical models working at different aggregation and logical levels, which interact to compute the most relevant performance indexes. The global computational effort is so low that it can be effectively used inside iterative optimization procedures for high speed networks.