Seven Ways to Flog the Data Center Energy Hog

By Thomas Fisher

When I was charged with the task of looking at the energy use of two data centers in the Boston metro area for Staples as an EDF Climate Corps fellow, two things came to mind immediately. The first was that I knew that data centers were unruly hogs of energy use and IT organizations had only just put together the standards of how to control them. The second was that I hadn’t ever stepped foot within a data center facility, and I needed to get my homework done well in advance to effectively tackle the project. Being an old efficiency consultant on projects ranging from financial services to the automotive manufacturing sector, I wasn’t too worried with applying the methodology I had learned in the past. The challenge this time was to tame the data center beast that had started to really plague the operations of Staples, and set them on the path to keep it in check going forward.

Taking a practical, holistic, systems thinking approach, I wanted to limit the scope of the project to what could be done with the data centers as they stand today versus what are some of the strategic discussions Staples should have around the future of their data centers overall. In the end, recommendations were given for short-term airflow, HVAC, monitoring and lighting fixes that would produce annual cost savings of $395,870, reduce the amount of kWh used by 2,839,467 and eliminate 2,038 metric tons of CO2. However, I still wanted to keep Staples focused on some of the best-in-class scenarios that could push them into innovating within the space in the long-term. Those seven recommendations included:

1) Ongoing virtualization and consolidation

Most organizations have already taken an initial pass at virtualizing existing set-ups over the past ten years. However, it is always important to remember that the low hanging fruit originally obtained from previous initiatives may have grown back over time, and there may be a need for periodic assessment to find where other larger opportunities still exist. Also, with the pace of change within the IT industry, equipment installed even a few years ago may have become outdated by more efficient and powerful systems that can result in dramatic reductions in space requirements and improvements in energy efficiency / use.

2) Joint IT / Facilities energy team formation

Being a native Midwesterner, I know that silos only work for grain. In most cases though, IT departments install equipment in their facilities without thinking of its corresponding energy use (which Facilities normally pays for). This creates a split incentive in which those who are most able to control the energy use of their IT equipment have little incentive to do so.

To gain optimum energy efficiency for those data centers, it is imperative to increase interdepartmental collaboration in the operations and monitoring of those facilities. Bringing the key stakeholders in IT and Facilities together (along with select others) to establish definitions, standards, goals, processes, procedures for procurement, etc. removes that split incentive and creates opportunities for effective communication and savings on all levels.

3) Enterprise wide energy efficiency budget allocations

Sometimes, energy efficiency projects get elbowed out by competing priorities within departments; stalling or curtailing the drive for organizational energy efficiency overall. Creating a capital relief program that specifically addresses energy efficiency projects as separate from normal budgeting allocations within business units would allow for smaller projects to proceed with the intention of increasing available funding for larger scale projects to tackle energy reduction goals. This would encourage a shift in thinking from shorter-term needs and projects with corresponding quick returns to those of the longer-term, strategic variety.

4) Implementation of comprehensive metrics

As the old mantra goes, “You can’t manage what you don’t measure.” For data center managers to understand how to effectively control the power and environmental conditions within their facilities, metrics must be ingrained in their decision-making processes. To start, a baseline of the existing available data must be conducted to figure out where the gaps are and what actions need to be done to fill them in. Using environmental or building management systems going forward to then monitor those key facility data points will enable stakeholders to continually make strategic decisions in the correct direction that mean the best results for power, environmental conditions and energy efficiency.

5) Distributed generation (DG)

Due to the steady power and cooling loads of most data centers, they are usually great candidates for the installation of fuel cell, solar or wind DG systems to produce power on site and take the facilities off the main grid. When used with combined heat and power (CHP) systems that use waste heat to provide cooling, the systems can produce attractive paybacks and reduce data center energy costs. With the various incentive programs available at the state and local levels to reduce the upfront cost of these installations, and the reduction in greenhouse gas emissions and pollutants that come with the use of clean DG, it is a solid choice for data center managers to explore in their acquisition of power for their facilities.

6) Say Hello to the Evergreen State!

In March of 2010, the Governor of Washington State signed into law a tax break provision for those companies that wish to relocate their data centers to a rural Washington county. The tax break allows for purchases of new data center equipment to incur no sales or use tax until 2018. Organizations would have to take advantage of the incentive program by July 2011, and the break only applies to facilities 100k square feet or larger, or spaces occupying at least 20k square feet within a 100k square feet or larger colocation facility. Additionally, existing colocation facilities in central Washington offer very favorable pricing for power (1.96 cents per kWh derived from 98% hydro and 2% wind) and leased space. Therefore, for those organizations that may have fragmented or underperforming facilities, long-term options such as relocation to Washington State may be worth the investigation for their overall data center strategy.

7) Containerize!

Since Google unveiled their container data center solution to the world in 2009, vendors have been building drop-in solutions for their customer base ever since. These solutions obtain best-in-class power usage effectiveness (PUE) and energy efficiency without the higher build out cost of establishing a containerized data center set up from scratch. With the lowest corresponding capital and operating expenditures in the industry, these solutions also have the quickest return on investment (ROI) and allow for incredible expansion flexibility with the growth of business. Hence, data center managers should seriously consider containerization options with their long-term strategy.

Now that my 10-week project is over, I’ve reaffirmed my beliefs in how easy it can be for organizations to find energy efficiency opportunities when the resources and bandwidth are there to do so. Indeed, small projects like this help organizations along in their sustainability initiatives by tackling the immediate issues the company is facing currently. However, when short and long-term strategy is addressed with a sustainable, systems-thinking approach, the growth, security and prosperity of that organization is more secured for the future overall.


By Thomas Fisher, EDF Climate Corps Fellow at Staples, Inc., MBA Candidate at Presidio Graduate School, Member of Net Impact.

3p Contributor

TriplePundit's guest author program has had over 600 contributors over time. If you'd like to be a guest author, please get in touch!