Optimal Mobility-Aware Handoff in Mobile Environments

Mobile terminals with multi-radio devices have become increasingly prevalent. This makes it possible for Internet applications being supported by heterogeneous wireless networks while the terminal is on move. As the user is constantly moving, it is highly desirable that the terminal connects to the best network and retains high performance of Internet connection. Handoff can be made within the same type of network or different types of networks. A vertical handoff may interrupt the ongoing activity but may also bring higher performance. Handoff in such a heterogeneous mobile environment faces several challenges, such as user mobility randomness, handoff overhead and optimality requirement. Existing work often focuses only on the current network condition when making handoff decisions, ignoring user mobility patterns. As a result, a handoff decision good for the present may soon become poor when the user moves to another place. Importantly, the current handoff decision may influence the future performance of the user´s connection. This paper proposes an optimal handoff algorithm that explicitly exploits user mobility patterns, with which we can effectively predict further locations of a user. This algorithm can produce optimal handoff decisions in the long run. Employing a comprehensive framework for preference customization, the algorithm supports user customization caring for different user preferences. We conduct extensive real trace driven simulation experiments and comparative study shows our algorithm is better than the typical algorithms.