Predicting multimedia traffic in wireless networks: A performance evaluation of cognitive techniques

Traffic engineering in networking is defined as the process that incorporates sophisticated methods in order to ensure optimization and high network performance. One of the most constructive tools employed by the traffic engineering concept is the traffic prediction. Having in mind the heterogeneous traffic patterns originated by various modern services and network platforms, the need of a robust, cognitive, and error-free prediction technique becomes even more pressing. This work focuses on the prediction concept as an autonomous, functional, and efficient process, where multiple cutting-edge methods are presented, modeled, and thoroughly assessed. To this purpose, real traffic traces have been captured, including multiple multimedia traffic flows, so as to comparatively assess widely used methods in terms of accuracy.