I was reading some articles online recently and I’m astounded by the lack of understanding of HACs (Hot aisle containment) versus CACs (cold aisle containment).
Let’s get right to the point:
1) Any containment is more efficient than no containment
2) If you are in an existing data center do HACs if you can – CACs may be more practical
3) If you are building new and/or doing a major upgrade always use HACs.
There, now I feel better.
It’s frustrating to me that this concept of HACs and CACs gets confused and combined with whether you choose to use a row or a room based cooling methodology. People seem to have hard time understanding that I can do row based cooling with HACs or CACs and I can do a room based cooling with HACs or CACs. The cooling methodology is independent of the containment.
So vendors who make most of their money from room based cooling systems should realize that they still have a play if a customer implements a HAC.
Here’s the kicker – in a theoretical world the efficiency of HACs and CACs is the same. The problem is that theoretical worlds only exist in engineering classes. The reality is that HACs is more predictable and practical in an operating data center for the following reasons:
1) It doesn’t require a raised floor for air distribution. Using a raised floor for air distribution, while the standard in the data centers for the past decades, introduces risk in terms of static pressure and consequently uneven air flow resulting in hot spots. I saw a Christmas tree stored under a raised floor once…that wasn’t helping air flow. All raised floors do is introduce systemic risk to your air distribution – because people do stupid things with them.
2) HACs allows you to get higher return air temperatures to the air handler. Higher is better and more efficient. I wrote our white paper #135 comparing the impact of cold aisle containment and hot aisle containment on the ambient environment. The reality is that with CACs you aren’t going to run the ambient environment as hot as you will the hot aisle with HACs. So you will, practically speaking, always get a higher return temperature with HACs.
3) Because HACs keeps the rest of the room cool so that you can have equipment in the uncontained space. I have yet to see an enterprise data center where they could contain everything.
4) You get less leakage with HACs adding to the efficiency gains. Raised floors leak (except in theoretical engineering classes). We have measured leakage rates of greater than 35% in real data centers. HACs is closer to 10-15%.
Maybe my mechanical engineering background or middle age makes me less tolerant of comments like this one I heard recently: ‘Containing the cold aisle means you can better control the flow and volume of cool air entering the front of the servers’
Don’t let this guy design your data center. The fact is that air will flow wherever there is low pressure. In a HACs system as long as there is enough CFM coming into the room, the cold air will find its way to the IT equipment. Unless you have lots of walls and obstructions inhibiting airflow (and I’m pretty sure I have yet to see a room that has more obstructions than a typical raised floor).
When people make comments like these I like to ask this question: If I build a CACs system utilizing the raised floor for air distribution, don’t I have the risk of static pressure and obstruction which would make it more difficult to control air flow? Particularly if the raised floor doubles as Christmas ornament storage?
That being said, I don’t want my position to be misunderstood. Any containment is better than none and CACs is a great strategy in an existing data center and I have recommended it to customers. Many times it is cheaper, easier, and faster to implement in an existing data center.
But it should be viewed as a tactical solution to a problem – not a long term strategy.
Please follow me on Twitter!
About Kevin Brown:
Kevin Brown is Senior Vice President of Innovation and Chief Technology Officer for the €3.7 billion IT Division at Schneider Electric. In this role, he is responsible for driving innovation and managing the R&D portfolio for the IT Division as well as driving the overall Schneider Electric portfolio strategy for the Data Center market. Prior to this position Kevin served as Vice President, Data Center Global Strategy and Technology. Kevin has also held numerous senior management roles in product development, engineering, and software development in the power electronics and HVAC industries. He holds a Bachelor’s of Science in Mechanical Engineering from Cornell University.