With summer in full swing here in the U.S., it seems a good time to deliver the third post in our series, Fundamentals of Cooling Architecture, which maps to three free courses currently available at Schneider Electric’s Energy University program.
The Fundamentals of Cooling I course looked at the three basic elements of a data center cooling architecture – heat removal, air distribution and location of cooling units – while the second examined humidity and how best to control it. This third installment is focused on choosing between room, row and rack-based cooling for data centers.
It’s a timely issue because more and more companies are installing high-density IT equipment that traditional data center cooling architectures are not designed to deal with. If data centers stick with such architectures, they often wind up with cooling systems that are oversized, inefficient and unpredictable.
As the course explains, room, row and rack-based cooling methods have been developed to address the issues that high-density IT racks and rows present. In a nutshell, the issue is that nearly all IT equipment takes in ambient air and ejects heated exhaust air. The air conditioning system must capture all this heated exhaust air – from potentially thousands of devices – and eject it from the room.
Perimeter, room-based cooling units have historically done an adequate job at this task. But as the course explains, simulation data and experience show they are only effective when the average power density in the data center is on the order of 1-2 kW per rack. But with modern IT equipment pushing peak power density to 20 kW per rack or more, a new approach is required.
The major difference between room, row, and rack-based cooling lies in how they perform the critical function of distributing air to the loads. The course compares and contrasts the various approaches and shows appropriate applications for each. Interestingly, you’ll find that many of them work with or without a raised floor (the topic of another blog post, based on a newly revised white paper titled “Raised Floors vs. Hard Floors for Data Center Applications.”)
As the course explains, nothing prevents you from using a mix of room, row and/or rack-based cooling systems. Such a hybrid approach makes sense in instances such as when you make upgrades that include installation of high-density racks where none existed previously in the data center. You may use row or rack based cooling for those racks while continuing to use the room system for the rest of the space.
You’ll also learn how to make an effective choice among cooling methods by comparing them against various criteria that are important in data centers, including agility, systems availability, total cost of ownership, serviceability, manageability and more.
Finally, the course walks you through a case study that shows how various cooling decisions play out in practice. It takes into account “first costs,” meaning the up-front cost of cooling equipment and piping, as well as ongoing electrical costs, reliability and performance of IT equipment.
It’s clear that data centers will be seeing increasingly higher densities in the years ahead and cooling architectures will have to adjust accordingly. Learn how to make the most effective decisions for your data center by spending less than an hour with the free course, Fundamentals of Cooling Architecture III. You’ll find it in the College of Data Centers at Schneider Electric Energy University.