Transfer Scheduling Schemes for Data-Intensive, Interactive Applications

This paper proposes a transfer scheduling scheme for data-intensive interactive applications, and new performance metrics that can quantify the usability of these applications. The scheme, called allowable data first (ADF), prioritizes those data whose transfer time will be within a certain period; it is evaluated using the performance metrics that are derived from observations of psychology: the allowable probability and the mean estimated time. The evaluation results show that ADF can achieve high allowable probability and small mean estimated time, and thus is an appropriate scheme for those applications.