Design Your Data Center for Today and Tomorrow!

(800) 646-9099

How to Ensure Your Data Center’s PUE Calculation is Accurate

By admin on August 30, 2016 in Main, Raritan
0
1

PUE calculationA key metric in determining the efficiency of your data center is power usage effectiveness (PUE). Understanding the PUE calculation – developed by The Green Grid, information consortium and global authority on resource efficiency – is essential for successful data center management.

PUE calculation involves dividing the amount of energy used to power the entire data center by the amount of energy drawn by the IT equipment. Ideally, those numbers should be as close as possible to get a PUE reading that is close to 1 (meaning that all of the power is being used by the IT equipment). For example, if your IT load is 2000 kW and your facility load is 3000 kW, your PUE is 1.5 (3000/2000 = 1.5). A PUE reading of 2.0 is about average, and a PUE reading of 3.0 is highly inefficient. Once you know your baseline PUE, you can measure the effectiveness of energy conservation initiatives and best practices to improve it.

Arriving at the numbers for PUE calculation can be complicated, however. Sometimes it’s difficult to tell whether certain devices should be included, or only partly counted, in the calculation. Some mixed-use facilities have power subsystems that are shared with functions outside the data center. Here are some steps you can take to make your calculations as accurate as possible.

Assign Categories

The first thing to decide is how to classify each of your data center subsystems. Divide everything into three categories: IT load, physical infrastructure, or subsystems that should be excluded. Some of these classifications will be easy. For example, servers, storage devices, and networking gear count as IT, while cooling systems, UPSs, transfer switches, and generators are physical infrastructure. Other systems – such as outdoor lighting or personnel workstations, that which do not directly contribute to data center function – are trickier to categorize. The important thing is to create a classification system that can be applied uniformly across locations.

Create Estimates for Shared Devices

Some power-drawing devices or subsystems are used for both data center and unrelated functions. For example, a UPS might provide power for servers and office equipment. Simply omitting such devices from the PUE calculation can lead to significant errors. Instead, get an estimate of the percentage of power that goes to data center functions. One way to do that is to temporarily shut down non-data center related loads on the device and measure the remaining power consumption. Estimates for devices that are difficult to measure are always better than omissions, to ensure a more accurate PUE calculation.

Use the Best Tools

To measure facility energy consumption, you can simply rely on your utility meter. Information about IT equipment energy use can be obtained from your UPSs or remote power panels (RPPs), but the best source is your rack power distribution units (PDUs). For example, Raritan has a wide range of intelligent PDUs that feature the ability to meter kWh, current, and voltage at the inlet, outlet, and PDU circuit breaker levels with billing-grade accuracy. They allow you to measure the power usage of specific devices. And intelligent rack PDUs can communicate with your data center infrastructure management (DCIM) software to perform PUE calculations automatically. Raritan’s full-colored rack PDUs make the power chains clearly visible, which can improve your equipment categorizations for PUE calculation. (That’s just one of the 5 Great Reasons to Use Colored Rack PDUs.)

Getting the most accurate PUE calculations will provide insights about your data center that can lead you to find ways to reduce your carbon footprint and save on energy costs. You can continue to use PUE calculations to show how the changes you implement contribute to greater energy efficiency in the data center.

About the Author

adminView all posts by admin

0 Comments

Add comment

Leave a Reply

Your email address will not be published. Required fields are marked *

*