Two years ago, the Green Grid established an awards program to highlight organizations that effectively improve the energy efficiency of their data centers. Through the program, we’ve found some excellent examples of companies that are making great strides in this area. In this post, I want to share one example: Hitachi Ltd., winner of our Grand Prix Award.
Hitachi is a great example of a company that successfully meets our award criteria, which includes:
- Use of quantitative evaluation criteria, such as power usage effectiveness (PUE)
- Presence of organization-wide strategic goals to reduce energy consumption
- A plan for concrete action to reduce energy
- Continuous energy reduction activities and quantitative assessment of the results
- Publishing of results and support of industry-wide improvement activities
We’ve found that award-winning organizations typically have three aspects in common: they consider energy efficiency a strategic corporate initiative; they leverage PUE metrics, continuously monitor metrics, and share results; and they take tangible steps toward improvement.
When Hitachi was challenged by a client to improve the energy efficiency of the client’s data center, the company set a lofty goal: reduce total data center energy consumption by 50% in 5 years, starting in 2007.
To achieve the goal, Hitachi took a three-pronged approach: visualization, assessment/analysis, and improvement/optimization.
Visualization involves collecting multidimensional data that applies to a data center environment. Hitachi monitored the temperature, humidity, and energy in the server room and installed power meters on all power distribution panels, including those in the server room. Then the company routinely patrolled for abnormalities in the server rack environment and facilities. It also checked the power status of the building management system that collects measurements throughout the network, using anomaly detection to monitor environment status.
The assessment/analysis prong involves analyzing the root cause of energy inefficiencies. Based on the results of visualization, the company periodically conducts evaluation/analysis meetings and examines the data it collects. Hitachi found it was effective to form a small work group of data center operators to conduct these analyses. It also uses computational fluid dynamics (CFD) simulations for environmental analysis. Through such analyses, Hitachi found issues such as an airflow shortage under the floor and an imbalance between hot and cold air that was causing hot spots in the data center.
The third prong, improvement/optimization, involves defining improvement goals and executing various actions based on problems found in the “assessment/analysis” phase. For example, to improve the air flow shortage under the floor, Hitachi organized the cable under the floor, installed a partition plate under the floor, relocated the free access panel, and changed the fan placement for its grill panel. To improve the mixing of hot and cold air, Hitachi installed a blanking panel, reviewed the grill position, and changed out an air conditioner.
As a result of its effort, Hitachi achieved a 2°C temperature improvement at the top of the data center rack. By addressing the hot spot issue, the company was able to decommission two air conditioners. But it’s not like the project is over. Rather, Hitachi continues to make various improvements, such as replacing old air conditioners with higher-efficiency units, and using water spray at external cooling units. Overall, the company is on track to meet its five-year goal of reducing energy consumption in the data center by 50%.
At the Green Grid, we encourage companies to share what they’ve learned about data center efficiency with others – and we’re grateful that Hitachi has done just that. But they are far from alone. To learn how other innovative companies are addressing the issue of data center efficiency, check out the our white paper number 40, “Energy Efficiency Measures In Japan: Case Studies.” We think it’ll help you make some sound decisions about how to improve the energy efficiency of your own data centers.