There are many ways to reduce energy usage in a data center.  There are a number of different approaches on both the mechanical and electrical side that can save significant amounts of energy.  Some of the approaches we have used include:
• The use of air side economizers
• The use of water side economizers
• Specifying the most energy efficient power and cooling equipment available
• Analyzing the energy efficiency of the equipment at the loads you plan to run them at
• The use of VFD’s on chillers, pumps and fans to vary air and water flow rates
• CFD modeling to ensure that the HVAC design chosen is as efficient as possible
• The use of natural gas co-gen plants
• The use of return grills in hot aisles and ducting CRAC unit returns into the plenum
The list of things you can do to improve the efficiency of a data center is a long one.  However, the one thing we can’t lose sight of is the need for reliability.  There must always be a balance between saving energy in a data center and the need to keep the data center on line.  Any design or operating decision that we make to save energy needs to be analyzed to see what impact it will have on reliability and redundancy.  As one data center manager said, “no one every got fired for using too much electricity, but people get fired all the time for data center outages.”  Keeping the data center on line is always going to be goal number one with any organization that has mission critical processing requirements. 
We believe the key to balancing these two seemingly conflicting priorities is to take a fresh look at data center designs.  We need to be open to concepts that could show significant energy savings, but need to be further developed to ensure that reliability concerns don’t get lost in the rush to save energy.  It’s very simple to cookie cutter data center designs and not give any real thought to new concepts.  Replicating the design in multiple locations with just subtle changes is a common engineering practice.  This is not always a bad thing.  Mature designs tend to be proven reliable over time and are often improved with incremental changes.  However, it doesn’t leave a lot of room for creative thinking.

A good friend of mine is an electrical engineer named Charles Kreiger.  Charles was one of the leading data center design engineers for 30 years.  He’s now 84 years old and semi retired.  I say semi because he still calls a number of us with new ideas for improvements in the way we design data centers. He is constantly questioning why we design things the way we do.   I am always amazed at how creative he is in his thinking. He’s currently working on providing uninterrupted cooling in data centers.  His thinking is that we did it 20 years ago for water cooled mainframes, why can’t we do it today for servers.  It’s that type of creative thinking that is sometimes lacking in the data center designs we see today.
 In balancing out the requirements for reliability and energy efficiency, we need to get back to the basics of what we are trying to do – provide the proper environment for the equipment we are supporting.  If you look at data center design from that standpoint, innovative design solutions should become much more common.  By being open to new concepts, I’m confident that we can achieve our goal, highly reliable and highly energy efficient data centers.  I’m also confident that you don’t have to sacrifice one to get the other.

About the Author: Ron Hughes, President of California Data Center Design Group has been involved in the design, construction and operation of data centers for over 20 years.  In the last 8 years alone, his company  has managed the design of over 2,500,000 square feet of data center space in the US, Europe, Asia, the Middle East and Mexico

Comments (0)

Subscribe to this comment's feed