Resource-Aware Component Scheduling Rejuvenation Model Using Lagrangian Relaxation

Abstract

As the time of system running increases, the performance of the system will decrease. This phenomenon is called software aging. Recently software rejuvenation as an efficient method to counteract software aging emerges. This method recovers the performance of the system through cleaning the internal state of system. Now component-oriented programming concept has been proposed and spreaded. And software system based on component emerges. These components with different function not only promote software reusability and self-management but also software reliability and availability. This paper demonstrates a software rejuvenation model based on resource-aware component optimized scheduling to realize software rejuvenation and avoid system crashing. In this paper we give the Lagrangian Relaxation component scheduling algorithm under single constrained condition and multi constrained condition. Then we test the algorithm and evaluate the model. Comparing system reboot, this rejuvenation model reduces MTTR (mean time to recover) very much. So the continuity of the service is promoted and the cost of downtime is decreased.