Power usage effectiveness (PUE) is a ratio that describes how efficiently a computer data center uses energy; specifically, how much energy is used by the computing equipment (in contrast to cooling and other overhead that supports the equipment).
PUE is the ratio of the total amount of energy used by a computer data center facility to the energy delivered to computing equipment. PUE is the inverse of data center infrastructure efficiency.
An ideal PUE is 1.0. Anything that isn't considered a computing device in a data center (e.g. lighting, cooling, etc.) falls into the category of facility energy consumption.
Survey respondents’ average annual PUE in 2022 was 1.55 (see Figure 1), which means that, in aggregate, their data centers expend 55% as much energy on cooling, power distribution and ancillary facility functions as on IT. This is consistent with the average PUE trend observed by Uptime in recent years — improvement slowed markedly in 2014, with only marginal gains since then.
davis.etal.2022.uptimeinstituteglobal (pg. 6)
New data center builds routinely outperform the average, achieving PUEs of 1.3 and below using facility designs and more advanced equipment that are optimized for lower energy use.
Limitations
shehabi.etal.2016.unitedstatesdata (pg. 57)
The limitations of PUE, the most commonly discussed metric of efficiency, are generally understood, but a key issue it that PUE only measures the efficiency of the building infrastructure supporting a given data center and indicates nothing about the efficiency of the IT equipment itself.
→ PUE indicates nothing about the efficiency of the IT equipment. A great PUE value can be almost useless if the IT equipment is highly inefficient. This also includes the wasted energy of idle servers.
Example:
In a data center with a PUE (power usage effectiveness) of 1.65, the IT equipment represents 61% of the power demand, whereas in a data center with a PUE of 1.1, it represents 91%.
— Lawrence, A., & Dietrich, J. M. (2022)
shehabi.etal.2016.unitedstatesdata (pg. 34)
The slower rate of efficiency improvement in PUE relative to IT equipment is partially due to the slower turnover rate of a data center’s infrastructure relative to the IT equipment. The opportunities to improve data center PUE increase with larger data centers that have the ability to develop better airflow management and employ more efficient cooling equipment or advanced cooling technologies such as liquid cooling. Consequently, smaller data centers are still being measured with PUE values greater than 2.037 while large hyperscale cloud data centers are beginning to record PUE value of 1.1 or less.
→ To increase the PUE you have to change the data center's infrastructure. The turnover rate of such a change is slower than changing the IT equipment. Only big data centers are able to make such an investment, e.g. to develop a better airflow management or employ more efficient cooling equipment.
davis.etal.2022.uptimeinstituteglobal (pg. 7)
Relying too heavily on PUE as the industry’s key efficiency metric may reduce operators’ motivation to pursue IT efficiency improvements.
If energy demand growth in the past decade has lagged the growth in IT demand, it is only partially the result of data center efficiency initiatives. The biggest improvements are the result of increased hardware efficiency (from processors, memory/storage and more), virtualization and the rise in consolidation activities. Despite this progress, major energy efficiency opportunities remain untapped. This may, in part, be due to a misplaced management focus on infrastructure and on key performance indicators (KPIs) such as PUE.
Today it is more important to increase IT efficiency and level of usage than improving PUE.
bashroush.lawrence.2020.puetacklingit (pg. 11)
For example, running the workload on 3-year-old kit, with a PUE of 2 but increasing the utilization from 5% to 25%, reduces the energy consumption by 62%.
Yuventi, J., & Mehdizadeh, R. (2013). A critical analysis of Power Usage Effectiveness and its use in communicating data center energy consumption. Energy and Buildings, 64, 90–94. https://doi.org/10.1016/j.enbuild.2013.04.015
