Improved Data Accesibility for Mobile Computers Through Predictive Methods

该项目将制定方法，通过改进数据复制，使移动的计算机能够及时获取不断变化的数据。当存储在移动的计算机上的文件的远程副本被更新时，系统必须平衡数据陈旧性与使用稀缺且昂贵的网络带宽来获取新版本。 本研究将调查的方法来自动预测何时在便携式计算机上复制的文件应与远程副本，哪些文件是最重要的协调，以及远程副本可能有最新的数据。 该项目将建立在预言家预测文件囤积系统的成功基础上。开发的预测方法将依赖于已经为其他目的保留的信息，加上少量新收集的信息。 模拟、分析建模和测量将用于确定合适的方法。该项目将产生一个工作原型系统。这项研究将提高便携式计算机用户的数据质量。 它将适用于其他类型的数据，如数据库片段或缓存的网页。

This project will develop methods to give mobile computers timely access to changing data through improved data replication. When remote replicas of files stored on a mobile computer are updated, the system must balance data staleness against using scarce and expensive network bandwidth to fetch the new version. This research will investigate methods to automatically predict when files replicated on a portable computer should be reconcilied with remote replicas, which files are of most importance to reconcile, and which remote replica is likely to have the freshest data. The project will build on the success of the Seer predictive file hoarding system. The predictive methods developed will rely on information already kept for other purposes, plus a modest amount of newly gathered information. Simulation, analytic modeling, and measurement will be used to determine suitable methods. The project will produce a working prototype system. This research will improve the quality of data for portable computer users. It will be applicable to other types of data, such as database fragments or cached Web pages.