An innovative approach for harmonising availability and energy efficiency in data centres: A case study

Abstract

A Data Centre (DC) as critical, energy-hungry infrastructure is dominated by two main driving forces: I) Availability and ii) Energy requirements. As a result of increasing energy costs, algorithms for optimising energy efficiency have been devised. However, these algorithms do not take availability into consideration. This paper aims to present a combination of Failure Mode, Effects and Criticality Analysis (FMECA)/Reliability, Availability, Maintainability (RAM)/Energy Analysis as an innovative approach for harmonising availability and energy efficiency in DC. Based on various measures defined by FMECA/RAM, corresponding availability and reliability are modelled and calculated. In parallel, potential energy saving measures are included in RAM simulation to quantify their influence on the availability and reliability of DC infrastructure. As a result, a set of the most promising optimisation measures is selected. Results show that some energy saving measures are highly correlated with availability. However, required data centre availability can be achieved with improved energy efficiency if the right set of optimisation measures is implemented. This approach guides DC managers to identify improvement potentials in terms of availability and energy efficiency, providing a reliable decision basis for future investments.