Mobile hotspot selection and offloading

Abstract

The adoption of smartphones, their ability to configure as mobile hotspots (WiFi access points), together with access to cellular networks, creates novel possibilities of heterogeneous network access for users. Specifically, a smartphone user may access a cell tower directly using the phone's 3G/4G connection or may connect to another smartphone in its vicinity that is con figured as a hotspot. Many prior works have looked at offloading the data traffic generated by cellular users to WiFi AP(s) in their vicinity. In these works the WiFi AP chosen for offloading a user, accesses the internet independently of the cellular network to which the user is connected. For example, the chosen WiFi AP could be in the user's home or in a cafe that may have connectivity to the internet via cable or fiber. In our work, we restrict ourselves to scenarios where internet access is available only via the cellular network. However, not every user may connect directly to it. Users in the network may be split into hotspots and clients. Hotspots are the users that connect directly to the cellular network and provide connectivity to the internet to other users by allowing them to connect to their WiFi interface. Clients do not access the cellular network directly. Instead each client connects to a hotspot. The optimization problem is to fi nd the split of hotspots and clients, and the association between clients and hotspots, that maximizes network throughput while ensuring that users get at least the throughput they get when all are directly connected to the cellular network. In this paper, we formulate the optimization problem, detail its characteristics, and propose a novel heuristic approach to split the network. We evaluate the gains in throughput achieved by using the approach for networks containing up to 100 nodes. Median gains of 1:5 are observed.