Mobility-driven wireless offloading system: architecture, scheduling and economic analysis

Abstract

Mobile data traffic is growing exponentially, as smartphones/pads equipped with high computing powers and diverse applications become popular. Immediate solutions to this explosive traffic growth problem are (1) upgrading the network to next-generation networks such as LTE (Long Term Evolution), and (2) installing more cell towers of smaller cell sizes (e.g., picocell, femtocell). Though these solutions are effective in increasing cellular capacity, they incur huge installation costs and network upgrade may not be effective in near future as the capacity of next-generation networks is getting close to theoretical limits. In this thesis, we build a mobility-driven wireless offloading system as a complementary solution to mobile data explosion, which effectively combines heterogeneous wireless networks (cellular networks, wireless local area networks, and peer tethering) for speedy, economical, and energy-efficient mobile access. In Chapter 2, we design delayed offloading architecture which exploits (i) traffic diversity in terms of delay tolerance and (ii) opportunistic contacts to WiFi access points (APs) from human mobility. We collect WiFi connectivity traces of 97 smartphone users for 2 weeks and analyze inter-connection times to an AP and connection times. Our trace-driven numerical analysis demonstrates that for an hour of delay tolerance, our offloading system achieves 48% of capacity saving and 20% of transfer energy saving. In Chapter 3, we study how much economic benefits can be generated due to delayed WiFi offloading by modeling the interaction between a single provider and users based on a two-stage sequential game. We first prove that WiFi offloading is economically beneficial for both the provider and users, through equilibrium analysis in our game-theoretic market model. We also conduct trace-driven numerical analysis to quantify the practical gain in terms of provider revenue and user surplus. In Chapter 4, we design collaborative tethering ar...