| The widespread availability of heterogeneous wireless networks (hetnets) presents a resource allocation challenge to network operators and administrators. Overlapping network coverage should be utilized to its fullest extent, providing users with a fair share of bandwidth while maximizing the efficient use of the operator's resources. Currently, network selection occurs locally at the mobile device and does not take into account factors such as the state of other networks that might be available in the device's location. The local decision made by the device can often result in underutilization of network resources and a degraded user experience. This type of selfish network selection might not result in optimal bandwidth allocation when compared to approaches that make use of a centralized resource controller. The decision making process behind the selection of these networks continues to be an open area of research, and a variety of algorithms have been proposed to solve this problem.;An over-the-top handover decision service treats each wireless access network in a hetnet as a black box, assuming detailed network topology and state information is unavailable to the handover decision algorithm. The algorithm then uses network data gathered empirically from users to provide them with a network selection service that considers the current conditions of available networks in a given location. This is a departure from past designs of vertical handover decision algorithms, which tend to approach the problem from the perspective of individual network operators. The wide range of radio access technologies operated by different network operators that are available to a device within a hetnet, coupled with the mobile data offload effort, is the primary motivator behind our novel choice in direction. This thesis documents the design and implementation of such an over-the-top vertical handover decision service. |
