Availability evaluation of system service hosted in private cloud computing through hierarchical modeling process.

Cloud computing provides an abstraction of the physical tiers, allowing a sense of infinite resources. However, the physical resources are not unlimited and need to be used more assertively. The challenge of cloud computing is to improve the use of resources without jeopardizing the availability of environments. Stochastic models can efficiently evaluate cloud computing systems, which is needed for proper capacity planning. This paper proposes an availability evaluation from a system hosted on a private cloud. To achieve this goal, we created hierarchical models to represent the studied environment. Sensitivity analysis is performed to identify the most influential parameters and components that must be compatible with improving system availability. A case study supports the demonstration of the accuracy and utility of our methodology. We propose structural changes in the environment using different redundancies in the components to obtain satisfactory results. Finally, we analyze scenarios regarding DC's redundancy. [ABSTRACT FROM AUTHOR]