Cooperation and decision-making in a wireless multi-provider setting

In this paper we investigate network design for a wireless service provider using two orthogonal technologies: a WAN technology with uniform spatial coverage and set of LAN access points each with limited coverage. We assume that the system is designed so that users (or their agents) independently and greedily select among the two options based on maximizing a specified utility function which may be a function of the quality of the wireless link, distance to the access points, and/or congestion on system resources. We focus on two complementary aspects of this problem. On the one hand we study system performance under such decision-making strategies. We show convergence of decision-making process to an equilibrium, and that a congestion-sensitive utility can provide substantial (300%) performance improvements over natural proximity-based criterion. On the other hand, we consider various problems associated with dimensioning typically expensive backhaul links, for the WAN and set of LAN hotspots. Our results show how to best jointly exploit technologies with different coverage scales so as to statistically multiplex spatial load fluctuations in order to reduce backhaul costs.