Technological convergence, high-density server deployment, and growing storage needs are putting new pressure on data centers to deliver peak performance around the clock.

Rising costs and impending regulations add to this pressure, compelling the telecommunications industry to adopt more energy-efficient practices in their data centers.

The Uptime Institute, for example, estimates data centers use between 8 percent and 30 percent of all the energy consumed by an enterprise.¹ While this consumption is costly for any business, it particularly adds up for service providers, which rely on data centers to store and circulate information to their customers.

Regulators are also noticing data centers’ impact on carbon-dioxide emissions. In fact, a joint study by the Uptime Institute and McKinsey and Company estimates data centers contribute to 0.3 percent of the world’s carbon-dioxide emissions. These emissions will surpass those of the airline industry by 2020, according to the study. ²

Understanding Energy Usage
As service providers begin to reevaluate their data center operations, the best place to start is with an energy audit. Many IT managers have no idea how much electricity is consumed by the data center since the electric bill bundles data center energy costs into the entire facility’s energy usage.

The Uptime Institute, for example, estimates data centers use between 8 percent and 30 percent of all the energy consumed by an enterprise.

Service providers can get a handle on their data center’s energy consumption by using monitoring and metering tools to calculate the total load. This information can help networking engineers identify inefficiencies and even verify the accuracy of their electric bills.

When considering data center energy consumption, it is equally as important to look above the rack at the overhead lighting. Lighting typically represents about 3 percent of data center energy use, and retrofitting can maximize savings.

Service providers can also cut energy costs by applying only the most critical IT loads to their UPS.

Cooling Priorities
A new energy allocation approach can help networking engineers improve cooling energy usage. Under the traditional watts-per-square-foot method, cooling spreads out consistently across the data center. Changing to a watt-density-per-rack approach helps maximize energy use by pinpointing the areas that need more cooling.

This analysis is particularly helpful for addressing inconsistencies that often appear when data center add-ons are deployed without first conducting an energy analysis. It also helps networking engineers plan for future high-density environments.

Importantly, this approach can reveal hot spots in the data center caused by high-density servers. Network engineers can realize greater cooling efficiencies by moving these servers closer to computer room air-conditioning units.

Hot spots can also be minimized by spreading out the rack density. In full racks, for example, a “chimney effect” can occur when servers lower in the racks receive more cool air than the servers on the top, resulting in greater failure rates for the servers located higher in the racks. By leaving empty rack space -- or reserving the top spaces for low loads -- networking engineers can better ensure their servers are cooled equally.

Another technique is to position the largest heat generating computer equipment in a concentrated area and apply precision cooling only to this area.

Make Way for the Air
Sufficient air circulation is critical to getting the most out of data center cooling. Small adjustments, such as spreading out the loads or separating racks by one or two floor tiles, can make cooling more effective. Air circulation can also be improved by moving cables out from under the raised floor to an overhead cable tray.

Networking engineers should also maximize their cooling energy by eliminating the mixture of hot and cool air. Airflow co-mingling frequently occurs:

• Between aisles and over the top of racks, as well as around corners.
• Within the rack as hot air from the equipment exhaust makes its way to the equipment intake.
• Through floor gaps and holes that are not directing cool air to equipment.

Recently, The American Society of Heating, Refrigerating and Air-Conditioning Engineers (ASHRAE) released new standards³ that raise maximum room temperatures to 80.6 degrees Fahrenheit, from 77 degrees, after finding most new machines can handle higher temperatures. Slight increases like this can add up to big savings on an energy bill.

When considering airflow improvements, it is also important to look around the data center and partition off any large, open or unused spaces. Use blanking panels to cover gaps between equipment in the front of cabinets. Also, arrange computer equipment efficiently in smaller areas to ensure energy dollars aren't wasted on non-computer areas.

Planning for Expansion
Consolidating equipment also gives networking engineers a sense of how much room they still have to accommodate expansion. As networking engineers plan for future needs, flexibility and scalability are critical.

Future readiness also means developing a solid migration path to 10G, 40G, and even 100G. Consider integrating a structured fiber optic cabling system with a top-of-rack, end-of-row server-switching deployment strategy. Obviously, fiber is also an option as large quantities of older copper cabling can impede airflow under the floors and within equipment cabinets.

Most data center experts know server virtualization and cloud computing are here to stay. These solutions offer real possibilities for reducing energy consumption. Virtualization, for example, gives multiple servers the ability to operate on the same hardware, thereby reducing power and cooling use, while saving space and costs in the data center.

Future Considerations
The telecommunications industry is striving for bigger bandwidths to accommodate today's explosion of information, yet all of this data needs to be housed somewhere. An efficient data center offers organizations cost savings and flexible room for growth.

When considering a data center update, suppliers can offer information about energy monitoring, new networking standards and other resources. Design and project management, strategic relationships with other suppliers, and logistic capabilities can help a data center get up and running with speed, intelligence and efficiency.

Service providers that get their data centers ready now will be positioned to provide the essential services that grow business, enhance competitive positioning, and meet energy regulations.

Endnotes
1. Uptime Institute. "Welcome to Green IT 2.0, the "Second Wave" - Now What Are You Going to Do About It?" 2010.
2. Uptime Institute and McKinsey & Co. "Revolutionizing Data Center Efficiency." 2008.
3. American Society of Heating, Refrigerating and Air-Conditioning Engineers. "2008 ASHRAE Environmental Guidelines for Datacom Equipment." 2008.

ASHRAE: www.ashrae.org
McKinsey & Co.: www.mckinsey.com
Uptime Institute: www.uptimeinstitute.org

About the Author
Karl Griffith is director, enterprise market for Graybar. He has more than 30 years of experience in technical sales and marketing with a strong technical background in structured cabling systems, communications infrastructure, data centers, physical security, building automation, fiber optic cable systems and networks. Karl is also a BICSI Registered Communication Distribution Designer. For more information email karl.Griffith@graybar.com, visit Graybar.com/data-centers or follow Karl on Twitter @Data_Center.

What’s your take on this subject? Leave a comment and get the conversation going.