Optimizing the total energy consumption and CO₂ emissions by distributing computational workload among worldwide dispersed data centers

Major internet service providers have built and are currently building the world's largest data centres (DCs), which has already resulted in significant global energy consumption. Energy saving measures, from chip to building level, have been introduced gradually in recent decades. However, there is further potential for savings by assessing the performance of different DCs on a wider scale and evaluating information technology (IT) workload distribution strategies among these DCs. This paper proposes a methodology to optimize the electricity consumption and CO₂ emissions by distributing IT workload across multiple imaginary DCs. The DCs are modelled and controlled in a virtual test environment based on a building energy simulation (BES) tool (TRNSYS). A controller tool (Matlab) is used to support testing and tuning of the optimization algorithm. A case study, consisting of the distribution of IT workload across four different types of data centers in multiple locations with different climate conditions, is presented. The case study will illustrate.