When talking about increasing energy efficiency in a data center, a lot of people stop listening because the issue of sacrificing uptime always rears up its head . Uptime should not even be a topic of discussion. It should be a given. Even in this day and age, many facilities just leave the thermostat at a low level because it’s a safe bet. But it’s not necessarily the right thing to do for the environment. What if we could help the environment AND ensure that your equipment is safe?
There are lots of things that you can do as an owner/operator to improve energy efficiency of existing data centers — without it costing you an arm and a leg. Mind you this isn’t really a big issue for when you’ve just started populating the data center… this is the kind of problem that evolves over days and months as different pieces of equipment are added or taken out of an existing room thereby shifting the thermal load within the room. Anyway, what can you do? You can create a scale virtual model of your data center (complete with servers, racks, CRAC units, cabling trays etc) and then use simulation to test various scenarios. The possible combinations are limited only to your imagination: use blanking plates or cooling curtains to divert air, move servers up/down within the rack, turn rack fans on/off, move new/old servers around the room, turn on/off various CRAC units etc. Because you are not using your physical equipment, you are not endangering your equipment nor do you have to build downtime into your schedule to test various scenarios. It’s a win-win situation no matter how you look at it.
KlingStubbins, an architectural, engineering, and planning services organization, routinely uses airflow simulation to create better and more efficient designs. Most recently, they used simulation to improve a server rack cabinet efficiency and thus improve the energy efficiency of the data centers that house them. The potential energy savings is significant for their clients because the racks in question are used in large numbers in some of their clients’ data centers.
Michael Schwarz, a Mechanical Engineer and an Associate at KlingStubbins evaluated 11 different scenarios to see which one offered the best solution to the challenge at hand. In the end he found that a combination of actions makes it possible to disconnect the cabinet fans while reducing all server inlet temperatures below existing values. Now I don’t know how long it would have taken him to reach the same conclusion without simulation but Mr. Schwarz said: “it would have been very expensive and time-consuming to investigate these design alternatives by building prototypes and evaluating their performance in a live data center.”
I found this case study fascinating because it contains detailed information about the various tested scenarios. If you’d like to learn more, then you can read this technical case study in its entirety here.
I hope you find it a fascinating read. And if you’d like to find out more about how to use this type of simulation at your facilities, please watch the on-demand presentation titled: Learn How to Reduce your Data Center Running Costs here.
Until next time,