Powering and cooling data centers takes up a lot of room in most companies’ budgets. However, there are some steps businesses can take to lower those costs. 

As we’ve written previously, budgets for data center design and management are getting tighter. Many companies are responding by putting less of a focus on creating redundancies, and accepting more risk that the infrastructure will go down.

And for most organizations, that lower risk is acceptable, according to a report from research firm DCD Intelligence. While IT departments may stress the need for building data centers that are highly resilient and have minimal expected down times, companies in the past may have overestimated their needs to avoid those risks.

However, adding more risk isn’t the only way to cut costs in the data center. Companies with tightening budgets can also take to reduce the amount they spend on power and cooling.

How much does it cost to run a data center?

Recent reports have made some shocking claims about how much energy — and therefore, money — companies spend on power and cooling. For example, one report in the New York Times claimed that a typical data center can waste 30 times as much power as it needs to run basic operations.

The main issue, according to the story, is that data centers often have a much higher capacity than they really need. In fact, the report says, the average data center runs at only 6% to 12% capacity, meaning just fractions of the data center power consumed is actually used to perform useful computations.

Those claims have been disputed — some observers pointed out that the research takes into account the data centers used by Internet giants like Amazon and Facebook, which have completely different needs than most other organizations.

However, while the exact amounts may be open for debate, one thing’s clear: Running a data center is expensive, and there are ways businesses can save.

3 steps to take

Companies can dramatically reduce power consumption in their data data centers with a few low-cost methods that can be applied quickly, according to a report from the IFMA Foundation.

The group recommends organizations:

1. Accurately monitor power consumption

Getting a clear read on how much power is being used has a couple of important functions. The first is that it allows businesses to find out what they can improve and to measure progress. Many businesses don’t have separate information about their data center, since it’s often lumped with their facilities as a whole. However, if power usage isn’t monitored, it can’t be managed.

The second reason monitoring is important is that many data center improvements require some level of investment, so it’s important for businesses to see how much their spending and to what degree they can reduce that figure. And of course, once improvements are made, it’s important to be able to tell if they’ve done what they were supposed to do.

The IFMA Foundation recommends businesses considering using dynamic monitoring — i.e., continuously monitoring power usage rather than taking an average over a period of time. That can help discover incidents that cause spikes in usage so that problems can be fixed.

2. Back off on data center cooling

Keeping data centers cool is important, but also takes up a lot of room in companies’ operating budgets. However, some small tweaks can lower the impact significantly.

One thing companies can do is to let things get a little warmer in the server room. Many companies keep their data centers cooled to around 70 degrees — however, last year Intel released a statement saying that customers could comfortable raise temperatures quite a bit without damaging servers.

Intel isn’t the first tech company to argue that data center temperatures are being kept unnecessarily low. In a 2010 presentation, Google’s “green energy czar,” Bill Weihl, said his company was keeping its server rooms at 80 degrees. And in 2011, Facebook announced a new data center in North Carolina would be kept at 85 degrees.

Experts say companies can check with their vendors to see the temperatures at which their equipment is certified to operate safely — and whether or not raising the temperature could void warranties.

3. Use scalable resources

One of the benefits of cloud computing is that companies can scale their capacity based on need, rather than building their own in-house systems with room to grow. That way, the organization isn’t spending money to power idling servers because they might be needed at some point. However, even with an in-house data center, companies can achieve some level of scalability, the IFMA Foundation says.

The group recommends companies develop “modular” data centers, in which the area is divided into separate zones whose power can be managed independently. That way, the company can scale up or down as needed, without having the entire data center powered on constantly.

Tags: cooling, energy, power, servers