What new data center metrics mean for cooling
Thursday, Dec 29th 2016

For years, metrics such as power usage effectiveness have been at the forefront of evaluating data center performance. PUE is calculated by dividing the amount of energy entering a data center by how much of it is needed to run the facility's various infrastructures. It is a ratio, with numbers closer to 1 demonstrating superior efficiency. The reciprocal of PUE is known as data center infrastructure efficiency; data center operators aim for a DCIE figure as close as possible to 100 percent.

"Many data centers still have a PUE over 2.0."

A 2013 survey from Digital Realty found that across the globe, many data centers still had a PUE over 2.0, meaning that they needed multiple watts of power to generate just a single watt of IT output. In 2014, the same company found that only 27 percent of respondents thought that their data centers currently measured PUE and many were unaware of the average. Over time, PUE has become less important to measuring the performance of a data center.

New metrics for the data center and how they affect cooling strategies

PUE has been around since 2006. Since that time, data centers have undergone significant changes in how they are designed and operated, in large part because of increasing demand. New metrics have emerged, such as data center energy productivity and performance indicator. The latter, which the Green Grid published in 2016, is particularly notable since it combines measurement of PUE with IT thermal compliance as well as resistance.

Many data centers are still inefficient when it comes to PUE and cooling.Many data centers are still inefficient when it comes to PUE and cooling.

PI allows data center operators to understand how reliably their equipment is being cooled. This adds another dimension to data center evaluation: Rather than just knowing how much incoming power is being lost, staff can also see how effective their cooling solutions are. For example, could a cable be moved to improve airflow and prevent heat buildup? Similarly, 3-D modeling of a data center floor, using computational fluid dynamics, is a useful way to identify points of potential cooling failures.

None of this is to say that PUE is obsolete, only that it cannot be analyzed in isolation anymore. Data centers have become more elaborate and essential than ever before, raising the stakes for safe and effective cooling. In addition to broadening the set of metrics they use, data center technicians should also invest in environmental monitoring solutions like the many products available from ITWatchDogs.

Ensuring effective cooling as data center metrics evolve

Your team can keep tabs on a wide range of conditions such as humidity, airflow, temperature and electricity usage by using web-enabled temperature sensors from ITWatchDogs. Notifications are easy to receive anywhere, via SMS, email, voice call and Simple Network Management Protocol. Find out more today about your options on our main product page.