Finite Horizon QoS Prediction of Reconfigurable Firm Real-Time Systems

Updating real-time system software is often needed in re- sponse to errors and added requirements to the software. Stopping a running application, updating the software, and then restarting the application is not suitable for systems with high availability requirements. On the other hand, dy- namically updating a system may increase the execution time of the tasks, thus, degrading the performance of the system. Degradation is not acceptable for performance- critical real-time systems as there are strict requirements on the performance. In this paper we present an approach that enables dynamic reconfiguration of a real-time system, where the performance of the system during a reconfigu- ration satisfies a given worst-case performance specifica- tion. Evaluation shows that the presented method is efficient in guaranteeing the worst-case performance of dynamically reconfigurable firm real-time systems.

Amirijoo, Mehdi, Aleksandra Tesanovic, Torgny Andersson, Jorgen Hansson, and Sang Son. "Finite Horizon QoS Prediction of Reconfigurable Firm Real-Time Systems." University of Virginia Dept. of Computer Science Tech Report (2006).