3G wireless core network dimensioning in the presence of self-similar traffics

Future all-IP 4G wireless networks will be based on the IP protocol and will be characterized by a 3G core network architecture. All-IP 4G systems supports multimedia traffics with heterogeneous characteristics. It is anticipated that the integrated traffics at the core network level exhibits long-range dependent behaviors and self-similar characteristics, as those in the current Internet backbone. This paper deals with an analytical approach for dimensioning networks with such traffics, modeled as fractional Brownian traffic. By imposing a total network cost, we derive the optimum capacity allocation minimizing an end-to-end delay metric. The results of our dimensioning technique (supported by simulations) highlight significant improvements with respect to the classical capacity allocation problem for Jackson networks.