Skip to Main content Skip to Navigation

Contrôle de qualité optimal d'applications multimédia

Abstract : We present a method of fine grain QoS control for multimedia applications. The method allows the control of application software whose actions are parameterized by a quality level parameter and whose execution times are unknown. The method allows the construction of a Controller which computes adequate action schedules and corresponding quality levels, so as to meet safety and optimality for a given platform. In other words, we want to have a maximal utilization of the available time budget without exceed it. The Controller uses a quality management policy for choosing, for each action, a schedule and quality levels meeting the QoS requirements. We extend and improve results of previous into two directions. We propose a symbolic quality management method using speed diagrams, a representation of the controlled system's dynamics. From this representation we propose a method allowing relaxing the number of call to the Controller with respect to the QOS requirements. In a second time, we developed a stochastic approach based on probability distribution functions for the execution time of actions. Our method is parameterized according to the importance attributed to deadline misses. This means that the user is given the possibility to express how hard the real time constraints are. Besides, given a value of the parameter, we can compute the expected deadline miss ratio for the controlled application. We present experimental results including the implementation of the method and benchmarks for a video MPEG4 encoder running on a bare machine.
Document type :
Complete list of metadatas
Contributor : Loïc Strus <>
Submitted on : Thursday, October 30, 2008 - 3:45:17 PM
Last modification on : Thursday, November 19, 2020 - 3:58:01 PM
Long-term archiving on: : Tuesday, October 9, 2012 - 2:42:41 PM


  • HAL Id : tel-00335812, version 1



Loïc Strus. Contrôle de qualité optimal d'applications multimédia. Modélisation et simulation. Université Joseph-Fourier - Grenoble I, 2008. Français. ⟨tel-00335812⟩



Record views


Files downloads