Traffic distribution schemes for multi-homed mobile hotspots

Providing broadband Internet access in mobile hotspots (e.g. public transport vehicles) through high-speed on-board local area networks and mobile routers is becoming an increasingly popular area of research and development. Mobile hotspot operators can provide faster, cheaper, and more stable communication services to on-board passengers using the multihoming technique, whereby the mobile router is connected to a diverse array of wireless access technologies (e.g., GPRS, UMTS, CDMA, and 802.11) through a multiplicity of wireless service providers. For mobile hotspot operators, how to "best" distribute the user traffic among the set of available access networks; which may change frequently during the trip; becomes a challenging task. In this paper, we propose and compare two traffic distribution schemes: in-transit and pre-transit, which are designed to take the benefit of multi-homing. The intransit scheme computes optimum traffic distribution on-the-fly, while the pre-transit scheme exploits the repetitive route of a public transport vehicle and computes the whole set of traffic (re)distributions before the trip starts. The objective of the traffic distribution computations is to optimize the profit and performance for the mobile hotspot operator. Using a generic public transport mobility model, we provide results from a detailed simulation study of these two schemes under different user demands.