How to Reduce Excess Cooling Capacity, as Inspired by Ken Brill8 min read
In the spirit of our founder, Ken Brill, and his “Gold Nuggets” of wisdom on data center efficiency, Upsite examines practical and effective steps to reduce excess cooling capacity.
By John Thornell, President, Upsite Technologies
Ken Brill’s Original ‘Gold Nuggets’ of Wisdom
Several years ago at an Uptime Institute Symposium, Ken Brill outlined five “gold nuggets” that could curb skyrocketing energy consumption in the data center:
1. Consolidating servers
2. Enabling server power-save features
3. Turning off servers no longer in use
4. Pruning bloated code to allow using less powerful servers
5. Improving the coefficient of data center energy efficiency
Ken was passionate about these gold nuggets because they were refreshingly simple, practical and effective – the proverbial low-hanging fruit.
Earlier this year (about a month before his passing) I was visiting Ken at his home in Maine. When we talked about results of research that sister-company Upsite had completed on cooling infrastructure, it sparked a passion that I hadn’t seen since the Symposium several years before. “I knew it! I knew it!” he exclaimed. “People are just throwing more cold air at their [IT] loads and still not fixing the root problem.” He was responding to Uptime Institute research that Upsite had recently updated.
The original study a decade ago had determined that sites had 60% bypass airflow*, and rated, running cooling capacity that was 2.6x the IT load. Coincidentally, this research led to the creation of both the KoldLok brush grommet and Upsite Technologies.
Upsite’s update to this research, with more sites and total square feet, determined that bypass airflow* had only been reduced to 48%. “That’s only about a percentage point per year over the last decade!” gasped Ken. He was right. We would have expected greater improvements, given advances in airflow management (AFM) and concerns about data center energy costs. More alarming than this, the average cooling capacity of the 45 sites researched had increased to 3.9x the IT load. Some sites had as much as 10x capacity. As Ken said, people seem to be throwing more expensive cold air at their sites instead of fixing the underlying AFM problems.
Updated ‘Gold Nuggets’ for AFM in 2014
In the spirit of Ken’s passion for data center improvement, especially simple, practical and effective improvements, Upsite offers the following AFM gold nuggets. These are designed to help data center managers determine their cooling utilization, identify AFM improvements, and then take steps to right-size cooling infrastructure to the IT load.
- Quantify current, excess cooling capacity using the Cooling Capacity Factor (CCF) metric.
- Make improvements in the right order using the 4’R’s (raised floor, rack, row, room).
◦Seal cable openings.
Immediate payback: Increased flow through perforated tiles & recovered stranded capacity.
◦Manage perforated tile locations.
Immediate payback: Increased flow through perforated tiles & recovered stranded capacity
◦Install blanking panels.
Immediate payback: Improved IT temperatures & better containment effectiveness.
◦Raise cooling unit temps.
Immediate payback: Reduced operating costs & recovered stranded capacity.
◦Reduce fan speeds and turn off cooling units.
Immediate payback: Reduced operating costs & recovered stranded capacity.
- Repeat the CCF and 4’R’s steps to achieve and maintain an optimal cooling capacity level.
At the end of each Symposium, Ken would plead with the audience, “Now, go home and do something!” In the same spirit, he would want us all to do something about excessive, expensive and wasteful cooling capacity. We hope Upsite’s resources can help:
Free Bypass Airflow White Paper
*More accurately, “Raised Floor Bypass Open Area” in today’s terminology.
Airflow Management Awareness Month
Free Informative webinars every Tuesday in June.
0 Comments