Optimization of Autonomic Manager Components in Service-Based Applications

Cloud Computing, as a distributed computing paradigm, consists in provisioning infrastructure, software, and platform resources as services. This paradigm is being increasingly used for the deployment and execution of service-based applications. To efficiently manage them according to the autonomic computing paradigm, service-based applications can be associated with autonomic manager (AM) components that monitor, analyze monitoring data, plan and execute configuration actions on them. Since an autonomic manager is composed of four basic components, then autonomic management of new applications lies on determining how many instance of each AM component to use in order to minimize the cost and maximize the management performance. In this paper, we propose both exact and approximate approaches that aim to choose the right number of AM components for managing service-based applications. Experiments shows the efficiency of our approach.