Data center infrastructure and its internal processing power consume significant amounts of energy, and inefficient data centers will consume even more power. Monitoring metrics such as Power Usage Effectiveness (PUE) helps benchmark data center efficiency while controlling costs. Organizations and data center managers can use this metric (measuring data center efficiency through energy consumption) to then reassess the impact of any changes made to the data center facilities. This helps reduce power consumption and energy costs.
PUE was created by members of The Green Grid, an industry organization focused on data center energy efficiency. According to the Uptime Institute's 2024 Global Data Center Survey, PUE and power consumption are among the most closely watched sustainability metrics. A similar benchmark recognized by The Green Grid is Data Center Infrastructure Efficiency (DCiE).
How is PUE calculated?
The following formula calculates power efficiency:
PUE = Total facility power / IT equipment energy consumption
Total facility power refers to the power consumption of the facility during operation, including all data center hardware, power supply components, cooling systems, and lighting systems. IT equipment energy consumption refers to the energy consumed to power on-site storage and network equipment, as well as control equipment such as monitors and workstations.
A PUE value must be higher than 1.0—ideally between 1.3 and 3.0—indicating efficient power delivery. Influencing factors include: using cabling infrastructure to minimize or eliminate power loss, and achieving equipment cooling without the need for fan ventilation.
DCiE is typically compared to Power Usage Effectiveness (PUE). It is the reciprocal of PUE because it divides the energy consumption of IT equipment by the total energy consumption of the data center. For example, a PUE of 1.0 means a DCiE value of 100% (i.e., 100% energy efficiency), while a PUE of 2.0 means a DCiE of 50%.
Calculate PUE using the following steps:
1. Measure the energy consumption of the facility's electricity meters. If the data center is located in a mixed-use facility or office building, only measure the electricity meters supplying power to the data center. If there are no separate meters, estimate the power consumption of the non-data center portions of the building and exclude them from the formula. However, a better approach is to use Data Center Infrastructure Management (DCIM) applications. Instead of looking at the meters for estimates, install additional sensors to collect real-time energy consumption data to measure PUE.
2. After power conversion, switching, and regulation are complete, measure the load on the IT equipment. The most useful measurement point is the power distribution unit supplying power to the computer and network equipment. This measurement should represent the total power delivered to the data center server racks.
3. Perform calculations. After collecting the measurements, divide the total power of the facility by the energy consumption of the equipment to calculate the PUE.
Advantages and limitations of PUE
The advantages of measuring power supply efficiency include:
PUE calculations are repeatable. This metric is designed to be used multiple times to measure changes in data center efficiency.
The Power Usage Effectiveness (PUE) value measures more efficient practices. This metric helps identify effective ways to reduce energy consumption.
DCIM software integration. The DCIM software can automatically calculate PUE in real time and integrate the data into reports.
The PUE metric can be competitive. Organizations with a good PUE ratio can use their score in marketing materials.
However, data center PUE also has some limitations:
Total facility power calculations are not always accurate. The accuracy of total facility power calculations is a major factor affecting PUE measurements. If an organization needs to estimate this figure using shared meters, the final calculation result is only an estimate.
PUE does not capture rack-level power consumption. Rack-level power consumption is not included in the overall PUE calculation, which may reduce the accuracy of PUE.
PUE is not a marketing metric. While businesses can use PUE in marketing, it is intended as a tool for improvement and evaluation. Marketers should not rely on PUE excessively simply because it is a widely known concept.
How to reduce PUE?
The following steps can help reduce the Power Usage Effectiveness (PUE) of a data center to close to 1.0:
Virtualized servers. Multiple virtual machines can run their respective workloads on a single hardware server, thereby reducing energy consumption and saving space.
Improve cooling systems. Data centers require cooling systems to prevent overheating. However, refrigerant-based cooling systems consume a great deal of electricity. Improving these systems or reducing a data center's reliance on them can help lower its Power Usage Effectiveness (PUE).
Optimize cold air production. For example, use naturally cooled outdoor air or heat exchangers instead of relying on refrigerants.
Replace inefficient hardware. The quality and performance of some hardware may degrade over time, so if a server or storage system is performing poorly, it should be replaced.
Use energy-efficient uninterruptible power supplies (UPS). Power distribution designs should incorporate UPS systems to improve efficiency. More efficient equipment and shorter transmission distances contribute to increased efficiency.
Use energy-efficient lighting. While lighting typically accounts for only a small fraction of energy consumption, it is still a simple way to reduce electricity and heat generation. Replacing fluorescent lamps with LEDs in motion sensors and lighting controllers helps reduce operating power consumption and heat generation.
Energy efficiency system manufacturer
PUE measurement technology is typically provided by vendors who manufacture systems used to inspect and analyze the overall efficiency of data center infrastructure. The following are the types of companies that provide PUE measurement and expertise:
Cooling system suppliers. These companies provide systems and expertise in advanced cooling technologies to reduce energy consumption.
Data center infrastructure management tools. DCIM systems are responsible for analyzing and managing numerous issues related to data center infrastructure.
IT equipment manufacturers. Companies like Dell and HPE provide energy-efficient servers and storage devices.
Consulting services. The Uptime Institute and Green Grid provide consulting services to improve PUE.
