The Green Grid (TGG) ( www.thegreengrid.org ) is a nonprofit consortium of industries with more than 175 member companies around the world. The key focus of TGG is data centers around the world. TGG is fast becoming a global authority on data centers, and when they speak about the efficiency of modern data centers, companies stop to listen. Recently TGG announced a published a case study of energy efficiency in Data Centers in Japan. This makes for interesting reading and there is much to learn from companies running data centers that may not be very efficient as yet. In this article, I have summarized the key learning points from these case studies.
In one case, a company decided to implement a 50% reduction in total data center energy consumption over five years. The company followed a three pronged approach consisting of visualization, assessment and analysis, and optimization. These were continuous processes and not a one-time effort, which resulted in major improvements in efficiency.
- Visualization – this involved the collection of multidimensional data about the data center itself. Temperature, humidity, energy utilization of individual components was measured. Abnormalities and excessive variations in these parameters were detected and studied.
- Analysis – root causes of inefficiencies were analyzed and it was discovered that there was an airflow shortage under the floor that was creating local hot spots. There was also an imbalance between hot and cold air.
- Optimization – the under floor cable layout was improved and other changes were made to improve the air flow.
The result of the optimization was that the temperature on top of the racks reduced by 2o C. Two air conditioners were made surplus and removed. Further improvements have been made by replacing old air conditioners with more efficient ones. The company has already improved the efficiency of its data centers and more improvements are underway.
In the case of another company, it was found that the awareness of power consumption was in itself a catalyst for continuous improvement in efficiency. The company examined in detail the energy efficiency in each data center. It spoke to operations staff and to management and build standard data collection methods.
The collection of data about energy utilization was automated to the extent possible. Results of the initial study showed temporary increase in use of captive power during electrical power outages and increase in air conditioning load during installation of new equipment. Having seen the savings accrue from a simple awareness of the consumption, the company is now expanding the breadth of its reporting, automating the reporting functions and increasing the frequency of reports which will give the company a better idea of the rate of increase or decrease in its consumption and allow corrective action to be taken earlier.
In a third case, a company built a data center in Kyushu in 2008. The company adopted a hot aisle containment design. The efficiency of this design improves with the power density level in the data center. Thus as the data center gets more load, the efficacy of hot air isolation will improve. The floor was raised to 1 m height which is not the norm followed by Japanese companies. Cabling under the floor was routed in such a way that it did not impede the air flow. Intake fans in the data center were controlled based on the temperature inside. As the temperature increased, the fan speed was increased as well. This further reduced power consumption. The hot air thrown out by the exhaust system is not allowed to go waste, but is used to heat up a large green house that grows a large variety of exotic fruits!
This company now uses external air for 10% of its data center cooling needs and has realized a reduction of 40% in its total consumption with all the methods listed above. In general, it has been found that energy efficiency tends to be higher in containerized data centers. Besides, they are erected faster, and have lower operating costs. Japan did not use containerized data centers earlier because of building standard laws and fire laws, but now these have changed and containerized data centers are beginning to appear in Japan as well.
Experience shows that data center efficiency is best ensured by following a synchronized approach with both IT equipment and facility equipment working together to a plan. Seasonal variations must also be taken into account and companies now routinely prepare different operating modes to cater for different seasons.
One important issue about running a data center is managing costs. Implementing some of the methods discussed in this paper has resulted in large savings that eventually go on to make the data center more competitive.
Be Part of Our Cloud Conversation
Our articles are written to provide you with tools and information to meet your IT and cloud solution needs. Join us on
Facebook and
Twitter.
About the Guest Author:
Sanjay Srivastava has been active in computing infrastructure and has participated in major projects on cloud computing, networking, VoIP and in creation of applications running over distributed databases. Due to a military background, his focus has always been on stability and availability of infrastructure. Sanjay was the Director of Information Technology in a major enterprise and managed the transition from legacy software to fully networked operations using private cloud infrastructure. He now writes extensively on cloud computing and networking and is about to move to his farm in Central India where he plans to use cloud computing and modern technology to improve the lives of rural folk in India.
