Simulating demand-driven server and service location in third generation mobile networks

Abstract

In this work we present a model for a 3G mobile network, emphasizing some of its main characteristics, such as provision of various types of services, including Voice, Video and Web; classification of users according to their mobility behavior in different geographical areas and during different times of day; varying demand generation patterns for each class of user; and exploration of the knowledge about the geographical location of users in order to improve the resource and information distribution over the network. A mobility simulator was implemented to generate a 24-hour demand in a typical metropolitan area and an integerprogramming problem was proposed to optimize the server and service allocation in a dynamic fashion. Since the problem turned out to be NP-Hard, a heuristic method was developed in order to solve it. The obtained results were near optimal. The proposed approach allows a personalized and demand driven server and service location in 3G networks. © Springer-Verlag 2003.