RUI/CSR-PDOS: Cooperative Prefetching for Mobile Devices

点击翻译按钮获取中文摘要

A fundamental challenge for many mobile computing applications is datamanagement. Data should be available anywhere at anytime. However,this task is particularly difficult to support with respect tomanaging a user's personal data (e.g., data files, email, and cachedweb pages) across a collection of personal devices (e.g., a laptop,PDA, and mobile phone). Two properties of the mobile environment makedata management particularly challenging: network disconnection anddevice resource constraints, particularly energy constraints. Thegoal of this work is to provide maximal availability and consistencyof personal data as well as maximal aggregate device lifetime for acollection of personal devices. To accomplish this goal, the projectdevelops a set of cooperative prefetching algorithms that enabledevices to retrieve relevant data when connected to the informationsource. Further, the algorithms efficiently distribute the energyburden of performing prefetching across the collection of devices.The evaluation investigates the performance of the algorithms in animplementation running on a collection of laptop computers and iPAQs.Two primary metrics are considered: the aggregate lifetime of thedevices and the hit rate of user requests for data. The overarchingimpacts of this work will be improved usability of mobile devices andapplications and maximization of mobile device lifetime. Further, thestudent participants in this project are undergraduate women fromMount Holyoke College, one of the nation's finest all-womenundergraduate institutions. Therefore, the project will also serve toimprove the overall participation of women in systems research.