Sensitivity Analysis of Data Center Electrical Infrastructures Considering Aspects of Dependability and Cost

Abstract

In the last years, the demand and use of data center have increased greatly due to the services offered by them. There must be a concern with data centers to allow the infrastructure to be available most of the time. Considering this need, this work has the objective of conducting a sensitivity analysis in the electrical infrastructure of the data center, thus finding the most sensitive equipment, taking into account the availability, reliability and cost. After obtaining the most sensitive equipment, it applies warm standby and hot standby replication techniques. The present work makes use of a hybrid modeling approach that considers the advantages of Stochastic Petri Nets (SPN) and Reliability Block Diagrams (RBD) to evaluate the dependability and cost of equipment acquisition.