What is PUE and how is it Calculated?

Published on August 1, 2023,

by Michael Wilson

In our modern, digitally driven world, organizations heavily rely on data centers to provide essential processing power for various applications and services. However, this tremendous capability doesn't come without a cost. Data centers consume a significant amount of energy, accounting for 2% of all electricity use in the US, and that figure is only growing. This raises a critical concern: efficiency.

The Importance of Data Center Efficiency

Inefficient data centers consume more energy than necessary, leading to high operational costs and a waste of valuable resources. Optimizing this efficiency is paramount for reducing energy consumption and maximizing the use of existing hardware. Key metrics such as PUE (Power Usage Effectiveness) and DCiE (Data Center Infrastructure Efficiency) play an essential role in monitoring and improving energy efficiency.

Understanding PUE

Power Usage Effectiveness (PUE) is a standard metric introduced by the Green Grid in 2006. It measures the total power consumption of a data center, helping in assessing its efficiency. Underutilized servers and energy-hungry hardware can drive up costs, and addressing these issues is vital. PUE serves as a benchmarking tool, allowing comparisons over time and helping in identifying the most effective energy-saving initiatives.

Calculating PUE

The PUE formula is:

An example of PUE calculation for a data center with 100,000 kilowatts of total energy and 55,000 kilowatts for IT equipment gives a PUE of 1.81. Lower PUE scores indicate higher efficiency. Data from the Uptime Institute shows that the average PUE is 1.59, signaling that there's room for improvement.

Measuring PUE

Measuring total facility power is best done at the utility meter or through a “shadow meter.” For IT equipment energy, power distribution units (PDUs) are typically used. Automation through sensors and Data Center Infrastructure Management (DCIM) software can ease this process, providing real-time data and a more detailed understanding of energy consumption.

.

The Connection between PUE and DCOI

PUE is a key metric mandated by the Data Center Optimization Initiative (DCOI), which aims at enhancing federal data center energy efficiency.

.

What Is DCiE?

DCiE (Data Center Infrastructure Efficiency) is the reciprocal of PUE, expressed as a percentage. It was also introduced by the Green Grid to promote efficiency.

Calculating DCiE

The DCiE formula is:

DCiE Formula

Efficiency levels based on PUE and DCiE:

PUE	DCiE	Efficiency
3.0	33%	Very inefficient
2.5	40%	Inefficient
2.0	50%	Average
1.5	67%	Efficient
1.2	83%	Very efficient

.

How PUE and DCiE Help Manage Costs

Regularly measuring PUE and DCiE provides insights into energy consumption across a data center’s infrastructure. Combined with DCIM software, operators can make informed decisions to reduce waste and costs, leading to a competitive advantage.

.

How to Lower PUE

Some ways to enhance efficiency include:

Replacing Inefficient Hardware: Upgrade to more energy-efficient components.
Improving Cooling Systems: Enhance cooling efficiency or utilize “free cooling” when possible.
Virtualizing Servers: Create virtual servers to reduce energy consumption and save space. Typically accomplished with Data Center Infrastructure Management (DCIM) Software
Integrating DCIM, BMS, and HVAC Systems: Integrating systems, from buildings to servers to applications, to provide greater efficiency, streamlined management and improved insight, unifying the data center ecosystem. Referred to as Integrated Data Center Management

Conclusion

Data centers are vital to modern organizations but come with substantial energy consumption. Metrics like PUE and DCiE are indispensable in assessing and improving efficiency. By embracing these metrics and taking practical steps, organizations can significantly reduce operating costs and make their data centers more environmentally friendly. Investing in proper tools and approaches is crucial for long-term success.