Data centers are huge consumers of energy. Think about the amount of information processing that is required to deliver millions of search queries in Google at the same time, or the power behind the mobile 3G networks that ensures that you can surf the Internet on the bus.
There’s no doubt that optimizing these data center for energy efficiency could lead to 1) reduced energy consumption and 2) lower costs. In this article we are going deeper into what makes a data center more efficient.
PUE (Power Usage Effectiveness)
One of the first things you should consider is to implement installations that measure the PUE. What exactly is this value? The formula can be defined as the following:
PUE = Total Energy Consumed / Energy Consumed by IT equipment
In other words, by taking the entire power consumption of the data center and dividing it only by the power consumed directly by the IT equipment, this formula allows us to see how efficient we can deliver power to the IT equipment.
The ideal value for a PUE would be 1, but as this is only an ideal value, systems will probably never reach it. This would be a data center that does not require cooling/heating or air-ventilation, but the equipment itself consumes all the power. However, some facilitates have achieved and monthly average of less than 1.10 (Google’s record is 1.09).
Why Energy Conservation is Important
Costs of measures that we use to conserve energy and improve energy conservation in a data center could well be earned back in terms of saved energy costs within the first year of implantation.
Costs is therefore one of the main motivators. Maybe a more important one is the larger picture and the climate crisis. Data centers across the globe make up a significant amount of the human carbon footprint, contributing to global warming. The two main motivators is exactly the same that answers the question “why is energy conservation at home important?”
Most Efficient Ways to Improve PUE
Air conditioning is one of the key places where efficiency usually can be improved. Cooling requires a lot of energy and by using thermal imaging to find so-called weak spots in the design of the heating/cooling system, we can eliminate parts of the power usage. Another way to do this is to use computational fluid dynamics.
Increasing operation temperatures (increasing the temperature where the computers operate) could lead to lower cooling expenses. Could this be done in a safe manner? Depends on your particular situation, but in many cases it can be done without lowering the lifetime of the components too much.