Energy Conservation in the Data Center: 12 Simple Steps

User Rating:

/ 0

Energy. Power. Whatever you want to call it, it’s a top concern of many data center managers and designers. (And yes, there’s a difference between energy and power, but that’s another matter.) As data centers consume more energy and the cost of that energy grows, the importance of increasing energy efficiency will also grow.

Ideally, you’d like to have a huge capital budget and the opportunity to build a brand new state-of-the-art data center with a PUE of a hair over 1.0, but chances are you won’t get that opportunity anytime soon—if every. But that doesn’t mean you can’t take some simple (and often relatively inexpensive) steps to increase your facility’s efficiency. The following are 12 ways you can cut your energy consumption without completely redesigning or rebuilding your existing facility. They’re not in any particular order, since some may be more applicable to certain data centers than others.

1. Select an energy efficiency metric and use it. One of the best ways to convince the C-suite that your plan for increasing energy efficiency in your data center is worth the cost is to provide a clear and accurate means of quantifying your success. A metric such as the ubiquitous PUE (power usage effectiveness) allows you to measure how effective your strategy is and to thus focus on those areas that provide the greatest return. And no, measuring efficiency isn’t the same as increasing efficiency, but it is a virtually indispensable part of the latter.

2. Virtualize IT infrastructure. This is a perennial favorite of companies looking to increase efficiency and cut costs. It may or may not fit under the heading of “simple steps” to greater efficiency, and it does carry its own capital costs, but given its popularity and the undesirability of the alternative—servers and other equipment that consume power but perform no useful work most of the time—you can’t afford to ignore virtualization.

3. Consolidate. If you don’t need it, get rid of it—or at least turn it off. Why clutter your data center with equipment that goes unused? It can obstruct airflow, create a disorderly appearance and atmosphere and, ultimately, cost you money, especially if you leave it plugged in or (even worse) turned on. Consolidation often goes hand in hand with virtualization: do more with less, and then get rid of what you don’t need (or, better yet, don’t buy it in the first place). Another term for this step is right-sizing. Whatever the name, offload the excess and save yourself time, space and energy.

4. Take advantage of free cooling when possible. Wherever your data center is located, you can exploit free cooling during some portion of the year. Free cooling, such as air-side economization, requires some infrastructure, but less than a conventional cooling solution. Of course, you’ll probably need some CRACs to keep things cool when outside air temperatures are too high (such as in the heat of the summer), but free cooling can quickly pay back the small capital investment in reduced energy consumption. Sometimes the best things in life really are free.

5. Increase your operating temperature. Excessive operating temperatures can damage sensitive electronic equipment, but you may not get what you pay for by running your data center at very cold temperatures. The American Society of Heating, Refrigerating and Air-Conditioning Engineers recently raised their recommended operating temperatures for electronic equipment, and many data centers are cutting energy consumption and cost by simply raising the thermostat. Besides, what good is making a server last for 20 years if you are probably going to replace it in 5? Of all these recommendations for improved energy efficiency, this may be the simplest: just don’t cool your facility as much.

6. Turn off unneeded equipment. This one is a virtual no-brainer. Why run power to a piece of equipment that you’re not using? Many electronic devices even draw some power when they’re turned off (assuming, of course, they’re still plugged into a live outlet), so unplugging them or cutting power to their connection site is an even better strategy. This step has no capital cost, other than the time it takes your staff to find and shut down or unplug unused equipment. The savings in operational expenses begin mounting immediately.

7. Control airflow in your facility. The temperature reading at a main thermostat doesn’t tell the whole story. Often, the problem is not your operating temperature, it’s localized hot spots that result from a lack of sufficient airflow. By minimizing such hot spots, you can generally run your facility at a higher temperature. Well-regarded strategies like hot/cold aisle containment can help, as can simpler steps like removing clutter and ensuring proper placement of ventilation tiles. Placing temperature sensors at key locations can give you a better picture of the heat contours of your data center. And if you’ve got some spare capital on hand, you might even consider computational fluid dynamics (CFD) analysis.

8. Maximize your power distribution efficiency. Power distribution infrastructure is like leaky pipes and fittings: the more cabling, connectors and intermittent devices you have between the power source and the equipment, the more you lose to waste. By minimizing the number of voltage changes and other inefficiencies in your distribution infrastructure, you can cut the power consumption and cost of operating your equipment.

9. Plan carefully for future growth. Consider building your data center infrastructure in a modular fashion. This approach allows you to gain the benefits of “right-sizing” while still meeting demand for IT resources. Whatever your approach, don’t just buy a bunch of equipment because your demand is growing: take careful stock of what you need to meet that demand. Too much equipment wastes capital and, if running, can waste power. A plan for growth makes good sense from an efficiency and business perspective.

10. Outsource to the cloud. Okay, there’s some debate about how much energy efficiency is gained by this approach. But the centralization of cloud computing, along with the amortization of equipment costs and sharing of resources among numerous customers, would seem to offer a more efficient alternative to every company running its own IT operation. Cloud computing may not be feasible as a “total solution,” but you may be able to exploit it as part of your overall IT strategy.

11. Increase your lighting efficiency. Consider using more-efficient lighting in your facility. If nothing else, consider a policy or automated system for turning off lights at the end of the day or when the data center is not in use. Computers don’t need light to work—only people do. Although the savings in this area may be small, they can add up over time.

12. Enable your staff to work remotely as much as possible. If employees are not in the facility, they’re not turning on lights, altering the thermostat and so on. When they can access systems remotely, you can also avoid some of the hazards of (even competent) human presence in your data center (many downtime incidents are the result of human error). This approach also saves fuel through fewer commuting miles.

Not every one of these steps will be feasible for every facility, and some data centers will have little to gain in some of these areas. These steps are simple, but you should still evaluate each one for its potential return in your case. And you need not start big: even small changes can help, and as you improve your efficiency (and demonstrate it with PUE data or some other metric), you will likely get more leeway from company management to implement more (and maybe more expensive) energy efficiency strategies.