Abstract. Over the past years, Autonomic Computing has become very popular, especially in scenarios of Cloud Computing, where there might be several autonomic loops aiming at turning each layer of the cloud stack more autonomous, adaptable and aware of the runtime environment. Nevertheless, due to conflicting objectives, non-synchronized autonomic loops may lead to global inconsistent states. For instance, in order to maintain its Quality of Service, an application provider might request more and more resources while the infrastructure provider, due to power shortage may be forced to reduce the resource provisioning. In this paper, we propose a generic model to deal with the synchronization and coordination of autonomic loops and how it can be applied in the context of Cloud Computing. We present some simulation results to show the scalability and feasibility of our proposal.
One of the main reasons for the wide adoption of Cloud Computing is the flexibility in which resources and software services are provisioned on demand through the concept of elasticity. Implementing elasticity to tackle varying workloads while optimizing infrastructures (e.g. utilization rate) and fulfilling applications' requirements on Quality of Service still remains an open issue and should be addressed by self-adaptation techniques able to manage complexity and dynamism. However, since Cloud systems are organized in different but dependent Cloud layers, self-management decisions taken in isolation in a certain layer may indirectly interfere with the decision taken by an other layer and globally affect the performance of the whole Cloud stack. Indeed, non-coordinated managers may lead to conflicting decisions and consequently to non-desired states. This paper proposes a framework for the coordination of multiple autonomic managers in Cloud systems. This framework introduces two kinds of managers: (i) one for each application; and (ii) another for the infrastructure. To tackle the problem of interferences between these Cloud autonomic managers, we propose a coordination protocol based on inter-manager events and actions along with synchronization mechanisms. The goal is to improve the synergy between layers in a loose-coupling manner. We evaluated the approach through an experimental scenario on Grid5000, a real physical infrastructure testbed. In this use case, we show that our framework improves the synergy between cloud systems while dealing with conflicting objectives and negative interferences.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.