What is the Best Measure of Data Center Efficiency?
Data center efficiency can be effectively measured by calculating the performance per rack. The performance per rack metric details analysis at the rack level including power consumption, rack density, and the physical data center footprint and combines it with performance metrics/targets. Ultimately, this kind of analysis yields more sustainable data center designs for the modern cloud era.
PUE focus has misdirected the fundamental point where data center efficiency begins: the compute infrastructure. Efficiency starts with the server itself and is best measured at the rack level. Efficiencies in both power usage and real-estate consumption are easily highlighted and compared at a rack level and scale up to the datacenter in a linear manner. Rack efficiency analysis offers operators and architects a valuable tool in reducing the carbon footprint of a data center. PUE is an important metric to understand how effectively the energy is used to power the compute infrastructure, however it can detract from the goal of increasing the efficiency of each individual rack.
Measuring data center efficiency begins with calculating the total amount of power used to deliver the level of compute performance required. Power consumption is tied closely to the overall physical data center footprint as well as its carbon footprint. Less power, less space, and less carbon equate to lower costs and greater sustainability.
The three elements of data center efficiency include: