Streaming Data Rate Prediction Model for Multimedia Service Workflow

Abstract

In this paper, we address the problem of reasonably arranging the streaming data rate so as to exploit better efficiency of the multimedia services in a workflow. We first focus on the fundamental models, sequence model and parallel model, to cover the essence of the problem. Other types of models could then be deduced by compositions of these two models. Based on the queuing theory, we get the video streaming data rate prediction model. Our evaluations show that this model could predict multimedia streaming data rate very well. This work is important because with this prediction, better allocation of system resource can be done.