Data Transfer Optimization For Mobile Cloud Computing

In recent years, mobile devices have more and more powerful processing capability and larger storage space, and the wireless networks provide higher bandwidth and larger coverage, which leads to the emergence of mobile cloud computing. However, the limited capacity of mobile device's battery and unstable wireless network environments seriously affect the user experience of mobile cloud applications. Data transfer optimization in mobile cloud computing becomes significant.In the mobile cloud computing, the applications are generally prefetching-friendly or delay-tolerating to adapt to dynamically changed network environment. The data transferring between the mobile devices and cloud is modeled as queueing process. Controlling the data transferring means deciding which data enters the corresponding queue and selecting the appropriate network interface to send these data in the queue. The energy consumption model and bandwidth model are consumed to perform quantitative analysis. The optimizing problem are finally defined as minimizing the energy consumption while satisfying the bandwidth demand of the applications. Lyapunov optimization and drift-plus-penalty are used to implement online optimization through minimizing the bounding value of the model. Data transferring process are controlled by auxiliary variable, selection, data admission control and queue updating.Mathematic analysis and extensive simulation have demonstrated the proposed method can select the network interface online to reduce the energy consumption while satisfying the demand of the application's bandwidth. And the system always work in stable status.