Analysis of the impact of different strategies of maintenance policies on the availability cooling and power infrastructures of a data center system using stochastic models

Abstract

Data centers are constantly growing, driving demands of new technologies such as cloud computing, e-commerce, which forced the availability of these systems 24 hours a day, 7 days a week under penalty of great losses. Our study aims to provide support for the analysis of the impact of different strategies of maintenance policies, both preventive and corrective, considering the parameters adopted in service level agreements on the stationary availability of infrastructures of a data center system using the mechanisms of RBD and SPN modeling. The cooling and power infrastructures were analyzed, considering a proposed communication infrastructure, in architectures with increasing levels of redundancy. These architectures were proposed based on the allocation of adaptations of redundancy mechanisms, which are widely used in systems projects, to the components of the referred infrastructures with the highest levels of importance for availability. In addition, a case study was developed to analyze the stationary availability resulting from different scenarios that consider different maintenance strategies from the infrastructures/architectures studied through the application of the proposed models.