A method of push-based communications scheduling for minimum latency time

It is desirable in data broadcast that, when data are updated, a server out puts the updated data as soon as possible in order to deliver the latest da ta to clients. When two or more data updates take place at the same time, t he server has to choose which data to output first. We first define the lat ency time, which is the elapsed time from the time data are updated to the time they are broadcast. Then, we propose methods to reduce mean latency ti me under the circumstance that each datum has various user request ratios a nd update probabilities. Furthermore, we discuss output data selecting algo rithms which can decrease the computational load of the server. (C) 2001 Sc ripta Technica