Sep 16 2013

10 Steps to Optimize Data Center Power

Improve data center energy consumption and get more tech with less juice using these steps.

There’s one sure way to take advantage of the latest and greatest techniques for energy efficiency in data centers today: Build a new facility from scratch. That’s one way to adhere to the U.S. Green Building Council’s LEED specifications and other industry guidelines for environmentally sound IT facilities.

Unfortunately, a “greenfield” approach is not often practical. This is especially true for IT managers who are challenged to maximize existing investments while still making their operations as efficient, reliable and cost-effective as possible.

Fortunately, there’s good news for those who can’t take a clean-slate approach to energy management. A wide variety of tools and techniques exist to give most any data center new levels of power and cooling efficiency.

Best of all, some of these options offer quick wins that produce fast returns in the form of cost savings and power reduction. Others require new, but not overly burdensome, investments that can be implemented over time to steadily increase in efficiency and make it easier for IT shops to justify new expenditures.

Here are 10 power steps CIOs can take today and in the near term to take advantage of the latest efficiency innovations and optimize energy usage in their IT operations.

No. 1: Build the Business Case

Before launching a new energy efficiency plan, IT managers should build a business case for new investments to garner C-level support and demonstrate the benefits the organization will see from its investments.

Cost savings will be among the most convincing justifications, and the U.S. Department of Energy’s Federal Energy Management Program provides a range of resources that can help make the case. The FEMP site offers buying advice for energy-efficient products and publishes a link to an office equipment cost calculator that IT administrators can download to help them estimate cost savings they’ll see when using ENERGY STAR-approved computers and office equipment.

No. 2: Audit Data Centers for Energy Efficiency

With the business goals clearly defined, IT managers are ready to audit their operations. The resulting data will help internal staff and solution providers understand the current state of power management in the facility, identify problem areas and point to new technologies and best practices that can have a positive impact.

“We bring in trained professionals to walk through the building, access what’s there and make recommendations,” says Jack Pouchet, vice president of business development for hyperscale solutions at Emerson Electric, a vendor of power management and cooling systems.

One common problem the pros will look for is so-called ghost servers, machines that once supported a needed application that over time has been moved or decommissioned but the hardware remains running. These energy hogs double down on inefficiency; they not only no longer provide a useful service, their outdated engineering consumes excessive amounts of electricity compared to the latest server designs.

“A best practice is to do an assessment periodically to make sure you’re not paying for any of these ghost machines,” says Charles King, principal analyst with the information technology industry analysis firm Pund-IT.

Some vendors of power and cooling products, such as Schneider Electric, also offer software designed to identify any obsolete servers still consuming energy so organizations can keep close tabs on their environments in between full walkthroughs. “The application looks at each device on the network to see if it’s pulling power and to determine whether it’s performing properly,” says Steve Carlini, global director of data center solution marketing at Schneider Electric.

No. 3: Implement Quick-Win Solutions

ASHRAE, a standards organization for the heating and cooling industry, has good news for power-conscious CIOs. Today’s servers can run at higher temperatures than what was considered safe in the past, which means organizations can spend less to cool their data centers.

“This is probably the fastest, cheapest and easiest way to improve the energy efficiency of the data center,” Carlini says.

Another economical measure is to implement containment systems that enclose aisles around server racks to prevent hot air from mixing with cold. This strategy is particularly important because some industry estimates gauge that 70 percent of the energy data centers consume goes to systems for cooling and managing air, notes Craig Watkins, product manager for rack and cooling solutions at Tripp Lite.

Cold-aisle containment systems are relatively easy to retrofit in existing data centers and can help organizations optimize the refrigerated air they produce to keep servers cool. They offer another noteworthy advantage: Their economical up-front costs, combined with their effectiveness in reducing cooling expenses, yield fast investment returns.

“Cold-aisle containment is low-hanging fruit for energy management. Organizations typically see a payback in about six months,” Pouchet says.

Hot-air containment systems are also an option. They require more retrofitting than cold-air solutions, but these specialized duct rack enclosures keep data centers cooler by directing hot air outdoors or into air-conditioning systems, Watkins explains.

Quick-win opportunities may also materialize with blanking panels, which enclose open spaces in racks that currently aren’t being used to house servers. Panels from Tripp Lite and others are designed for the IT staff to easily plug into empty spaces without the need for special tools, which simplifies installation.

Finally, IT managers shouldn’t overlook the most basic energy economizer of all: untangling the balls of cables that build up around racks and other equipment that keep air from flowing efficiently.

No. 4: Upgrade Equipment to the Latest Energy-Saving Designs

Purchasing computers, servers, displays and other office equipment with an ENERGY STAR certification can reduce power consumption and help organizations save significant money. For example, program officials estimate that over a five-year timeframe organizations can save $500 per server because certified equipment averages 30 percent higher efficiency rates than standard gear. If all servers sold in the United States met the specification, energy cost savings would approach $800 million per year, ENERGY STAR experts say.

IT managers can adopt additional optimizations by shopping for the latest server designs that draw power dynamically, based on prevailing workloads. “In the past, servers used the same amount of power if they were sitting idle or running at high capacity,” Carlini notes. “Today’s energy-efficient models can reduce their energy consumption thanks to internal power management features.”

Buying the latest servers also will produce a ripple effect. Pouchet says servers have demonstrated significant gains in compute performance with each new generation, which means that running servers three or more years old creates unnecessary costs.

“You’ve got to know the age of the equipment you have and when to get rid of it,” he says. Saving one watt of electricity by implementing a more efficient processor results in cumulative savings in power distribution, cooling and other costs of approximately 2.84 watts of total consumption, he explains.

Energy-saving modernization opportunities exist for other types of data center equipment, too. Watkins advises organizations to modernize older uninterruptible power systems with energy-saving devices that run more efficiently and produce less heat, in turn reducing the strain on cooling systems.

No. 5: Take Advantage of Close-Coupled Cooling

The concept is simple: For best results, IT managers should place the cooling source as close as possible to equipment that needs protection from overheating. It’s an alternative to spending excessively to keep the whole data center as cold as a meat locker.

Carlini advises combining this approach with environmental sensors in server racks that monitor how hard servers are working. “So when the heat goes up, the fan speeds and the output from the cooling devices increase, too,” he says.

Look for self-contained, close-coupled cooling systems that the IT department can install without having to hire contractors and don’t require new plumbing, air ducts or other expensive alterations, Watkins says.

No. 6: Be Safe About Using Outside Air

It’s tempting to pipe in cool air from the outdoors to reduce the load on cooling systems, but experts advise data center managers to tread carefully. Industry insiders say that poor air quality, from high humidity levels and pollution, can cause problems for sensitive computing equipment if the air isn’t filtered properly.

A UPS with an Energy Star certification can reduce energy losses by 30% to 55%, which means that a 1000kVA UPS in a large data center could save $18,000 annually

“If there is pollution or a brush fire at a distant location, contaminants can be carried downstream and cause problems in data centers,” Pouchet explains. “It’s best to install ion sensors that can shut air dampers to reduce the flow of outside air if there are high levels of contaminants.” Filtering screens and humidifiers are additional ways of assuring outside air is safe for IT operations, he adds.

No. 7: Send Power and Cooling Challenges to the Cloud

Offloading workloads to cloud providers means CIOs can reduce their on-premises computing requirements and the energy demands that go along with them. This helps explain why 62 percent of IT professionals identified cloud computing is a factor in energy efficiency in the CDW 2012 Energy Efficient IT Report. Only 47 percent of the respondents made that connection in the previous report.

“Cloud data center are highly virtualized environments and cloud providers try to get the biggest bang from their energy investments,” analyst King notes. “So there’s a likelihood that energy is being managed more efficiently than in many traditional data centers.”

No. 8: Look for Power-Saving Opportunities Outside the Data Center

When shopping for PCs, notebooks, printers and other essential hardware, IT managers should look for energy saving innovations that can help reduce power consumption throughout the enterprise. For example, computers that boot up quickly, such as those with solid-state hard drivers, enable users to shut them down during inactivity rather than keeping them running and drawing power.

For other types of computers, install power management software that can put devices in low-power sleep mode to save energy when appropriate. Similar innovations apply to other types of equipment, such as multifunction and standalone printers. IT managers should confirm that power-management features are engaged in these units to reduce power draws between print jobs.

No. 9: Drill into Detailed Data

Data center infrastructure management (DCIM) solutions help IT administrators closely manage energy usage by gathering essential information about power and cooling systems and measure energy consumption by individual computing devices. These solutions, such as Emerson Electric’s Trellis platform, help organizations proactively manage energy usage and automatically enforce internal policies designed to reduce operational costs.

DCIM is poised for solid growth in the years ahead, according to IT market researcher IDC. In 2011, the market for DCIM software and services totaled $247.1 million. By 2016, the DCIM market will grow to $690 million, for an annual growth rate of 22.8 percent.

No. 10: Educate End Users

Vigilant IT and facilities managers can use social media, blogs, e-newsletters and other readily accessible tools to encourage co-workers to conserve energy. Turning off unused equipment and unnecessary lighting throughout an enterprise can provide a money-saving complement to the formal strategies CIOs employ to keep data centers running as efficiently as possible.

Wavebreak Media/ThinkStockPhotos

aaa 1