Tips to Manage and Drive Energy Efficiency for Every Data Center
It’s no secret that the deluge of digital content being created, shared and retrieved on a daily basis is having a direct effect on data center growth. Tens of thousands of data centers have been built to support this blast of digital data. Unsurprisingly, this has directly increased the use of electricity by data centers. Consider these facts:
- In his Stanford University study, Growth in data center electricity use 2005 to 2010, consulting professor and researcher Jonathan Koomey states on a worldwide level data centers used 56 percent more electricity in 2010 than in 2005. In the U.S., data center electricity use increased by about 36 percent for the same time period.
- This same study ascertains that data centers likely accounted for between 1.1 percent and 1.5 percent of total global electricity use in 2010. In the U.S., that number was between 1.7 percent and 2.2 percent.
- According to the Schneider Electric whitepaper, An Improved Architecture for High-Efficiency, High-Density Data Centers, less than half the electricity used in a typical data center actually makes it to the computer loads. More than half the electrical bill goes to the purchase of power consumed by the electrical power system, the cooling system, and lighting.
These facts stress the massive importance of data center energy management and the responsibility of data center owners and managers to implement and maintain energy efficient technologies to minimize environmental impact on a regular basis.
Benefits of Improving Data Center Energy Efficiency
By implementing environmentally-efficient solutions, such as installing data center infrastructure management (DCIM) software, looking at ways to reduce IT load or powering off unused equipment (like servers), executives are able to not only lower their environmental impact, but also make their businesses more profitable with simple usage cuts. By tackling issues including oversized power and cooling equipment and poor data center configuration that cause inefficiency, executives can save up to 30 percent in power consumption as compared to traditional designs. Improvements on the architecture of existing data center infrastructure, like taking advantage of modular infrastructure and installing high-efficiency uninterruptable power supply (UPS) systems, also result in improvements in not only efficiency, but capacity and reliability. Lastly, increasing data center energy efficiency also provides an effective response to increasing public pressure to reduce environmental impact.
However, effective energy management doesn’t have to come strictly by increasing the budget, or by employing major data center overhauls. Both new and legacy data centers, regardless of size, can take advantage of several low-cost methods to improve energy efficiency.
What causes inefficient data centers?
To improve data center efficiency, it’s necessary to understand what reduces efficiency and eats up power in the first place. These factors include:
- Power equipment inefficiencies with UPS systems, transformers, transfer switches, and wiring. Efficiency is drastically affected when equipment is doubled for redundancy or operated well below its rated power. Additionally, the heat generated by power equipment must be cooled by the cooling system, using up even more energy.
- Cooling equipment inefficiencies with air handlers, chillers, cooling towers, condensers, pumps, and dry coolers, it’s important to operate the equipment in a variable manner and closer to its optimum power usage curve. This is why adding variable capabilities through variable frequency drives (VFDs), electronically commuted (EC) motors, and automation systems are vital to an efficient cooling system. This allows the cooling supply to limit excess waste by more closely matching the IT (heating) load as it changes throughout the day, week, and year. Also, as with power equipment, when cooling equipment is doubled for redundancy or operated well below its rated power, efficiency falls dramatically.
- Lighting power consumption uses a large amount of power and generates heat. As with power equipment, the heat generated by lighting systems must also be cooled using cooling equipment. High-efficiency lamps and ballasts, as well as motion sensors, can drastically reduce the amount of heat generated, as well as provide energy savings. Simply turning off the lights in unutilized areas of the data center or during periods without the presence of personnel can also increase efficiency.
- Power and cooling system over-sizing is one of the largest drivers of electrical waste. This occurs when the design value of the power or cooling system exceeds the IT load, often due to the expectation of fast IT load growth. For example, the IT load may have been overestimated and the power and cooling systems were sized for too large a load. Or, the cooling system itself may have been poorly designed, requiring over-sizing of cooling equipment to successfully cool the IT load.
- Inefficiencies within the physical configuration of IT equipment, which can force the cooling system to move much more air that necessary, cause the cooling system to generate cooler air than the IT equipment requires, and/or force cooling units into conflict when one is dehumidifying while another is humidifying.
Cost Effective Methods for Improving Data Center Energy Efficiency
Knowing the importance of increasing data center efficiency and the reasons behind inefficient energy use, there are several tactics data center managers can implement to improve overall energy efficiency. Many of these tactics are low-cost and require very little disruption to daily operations.
Power off unused equipment. This is perhaps the cheapest, easiest way to reduce energy usage, and facility managers often fail to acknowledge how much electricity unused equipment is wasting simply by being powered “on.” Just as we’re taught to turn on the lights or switch off the television in a room we’re not currently occupying, it’s important to make sure only the IT equipment being used is turned on. Virtualization makes this especially easy. With the appropriate DCIM tools in place, managers can shift whole IT loads around according to time of day, and shut down entire lots of servers for hours at a time.
Optimize the physical data center layout for maximum efficiency. An example of this is the implementation of hot and cold aisles through your server positions. Physically separating hot and cold air through items like blanking panels, curtains, and containment decreases the need for unnecessary cooling capacity and reduces overall energy usage. Similarly, in-row cooling, which locates air conditioning within IT equipment rows instead of the perimeter of the room, shortens the path cool air has to travel and also reduces the chances of hot and cold air mixing.
Install high-efficiency UPS systems. According to a study by Lawrence Berkley National Labs, at 30 percent load, new, high-efficiency UPS systems are 10 percent more efficient when compared to the average UPS system. Actual wattage losses of high-efficiency UPS systems can be reduced by as much as 65 percent.
Implement a DCIM solution. DCIM software is a great capacity management tool that can help managers improve power utilization, cooling and rack capacity by monitoring energy consumption. DCIM software can identify and alert managers about specific conditions that create less than optimal electrical consumption, helping to identify opportunities for reduced electricity usage and improved energy efficiency.
Take advantage of scalable power and cooling to avoid over-sizing. Data center efficiency decreases when IT loads decrease. Therefore, when the percentage of IT load actually being used is below the design value of a data center, the data center is considered to be over-sized for that IT load. This particularly affects power and cooling efficiency. To correct this problem, power and cooling equipment should be scalable to meet data center needs over time. Not only does this increase efficiency, but also can help managers reduce capital investments and operating costs until additional capacity is required.
Consider modular, containerized solutions. In addition to tying up fewer resources – including upfront capital expenses, ongoing operational costs, space and manpower – modular, containerized solutions are pre-programmed to run as efficiently as possible. Modular solutions are packaged, shipped and installed as a container that plugs into an existing space, allowing for rapid increases in efficiencies to existing data centers.
What’s next for data center energy efficiency?
While not necessarily practical for the average data center now, large enterprise data centers and cloud and colocation facilities are at work implementing advanced technologies to improve efficiency and reduce environmental impact. This includes (but is not limited to) the implementation of free cooling methods, sourcing renewable energy, installing custom, high-efficiency servers and taking advantage of water-based cooling methods. These methods will continue to be refined, and may eventually be able to be scaled down for the average data center, possibly even as a cost effective retrofit.
Regardless of the size and type of facility, every data center has the responsibility to reduce energy usage by increasing efficiency.
Lance Bishop is director of energy management services for Schneider Electric.
- Let's Do The Math for DR
- 2014 Environmental Leader Product and Project Awards
- Improve Your Company's Environment and Energy Performance
- Gartner Magic Quadrant
- Unlocking the Value of Energy & Operational Data
- Smart Companies Utilize Integrated Energy Solutions
- Best Practices in Electricity Procurement
- How "Fixed" is the Fixed Price Product?
- Increase the Value of Demand Response Through Automation
- The Future of Operational Risk Management: The Oil & Gas and Chemicals Approach
- Connected Buildings, Connected People: A Look to the Future
- Cut Costs and Improve Facility Operations with Energy Data
- Energy Procurement Strategies for Winter 2014 and 2015
- Energy Efficiency Requires Engineering Efficiency
- Integrated Building Optimization: A Crucial Convergence of Demand-side and Supply-Side Energy Management Strategies