Sustainability Analysis in Data Center Dense Architectures

Abstract

Nowadays, energy projects for data centers need to consider the high consumption loads per rack due to the density achieved with modern IT equipment. Capacity management aims to ensure availability when supplying electricity at a safe level so that the available capacity is greater than the capacity used, without prejudice to future sizing, without burden to the total cost of ownership, but efficient enough to meet current needs. Nevertheless, organizations must include $CO_{2}$ emissions in their list of concerns because of electricity production/consumption. In this way, this paper estimates the critical load demand for data centers and propose energy flow models (EFM) for four architectures, evaluating the minimum energy required for the project and the metrics of exergy, costs, PUE, and efficiency. We use a Power Load Distribution Algorithm (PLDA) to evaluate energy consumption and calculate the impact of environmental sustainability concerning raw material used in power production. The results show that even a data center denser in IT infrastructure can emit less $CO_{2}$ and be more ecologically sustainable when the choice of energy sources is cleaner.