Kalinahia: Considering Quality of Service to Design and Execute Distributed Multimedia Applications

Abstract

One of the current challenges of Information Systems is to ensure semi-structured data transmission, such as multimedia data, in a distributed and pervasive environment. Information Sytems must then guarantee users a quality of service ensuring data accessibility whatever the hardware and network conditions may be. They must also guarantee information coherence and particularly intelligibility that imposes a personalization of the service. Within this framework, we propose a design method based on original models of multimedia applications and quality of service. We also define a supervision platform Kalinahia using a user centered heuristic allowing us to define at any moment which configuration of software components constitutes the best answers to users' wishes in terms of service.

Kalimucho: Contextual Deployment for QoS Management

“…We can represent applications as configuration graphs [10] where one application can be realized by one or several configurations with different QoS. For each application, configurations are classified according their decreasing utility.…”

A model-driven method for describing and predicting the reliability of composite services

“…Approaches that explicitly deal with the reliability prediction are proposed in [16][17][18]36,37,41,47]. In [18] an UML-based framework to predict the reliability of a two-tier application is proposed.…”

“…Then a refined testing is applied over a smaller set of candidates. In [17] a middleware model to allow the dynamic adaption of distributed multimedia applications that provide (and maintain) an adequate QoS level has been proposed. The execution platform that integrates both static and dynamic reliability prediction acts a reconfiguration of the distributed application by varying its composition, in order to provide the best QoS level.…”

Risk-based adaptive scheduling in randomly deployed video sensor networks for critical surveillance applications

“…The final goal is to provide the necessary algorithmic support for a surveillance application to express its objectives. How the objectives are translated into the deployed sensor network or how the application can dynamically configure nodes in the networks are out of the scope of this paper and may need some kind of supervision platforms that are proposed in the recent research literature (Laplace et al, 2008).…”