Free Cooling: the Server Side of the Story
by Johan De Gelas on February 11, 2014 7:00 AM EST- Posted in
- Cloud Computing
- IT Computing
- Intel
- Xeon
- Ivy Bridge EP
- Supermicro
Performance?
Yes, we did monitor performance. But it simply was not worth talking about: the results at 20°C inlet are almost identical to those at 40°C inlet. The only difference that lower temperatures could make is a slight increase in the amount of time spent at higher Turbo Boost frequencies, but we could not measure any significant difference. The reason is of course that some of our VMs are also somewhat disk intensive.
Conclusion
The PUE optimized servers can sustain up to 40°C inlet temperature without a tangible increase in power consumption. It may not seem spectacular but it definitely is. The "PUE optimized" servers are simply improved versions; they do not need any expensive technology to sustain high inlet temperatures. As a result, the Supermicro Superserver 6027R-73DARF cost is around $1300.
That means that even an older data center can save a massive amount of money by simply making sure that some sections only contain servers that can cope with higher inlet temperatures. An investment in air-side or water-side economizers could result in very large OPEX savings.
Reliability was beyond the scope of this article and the budget of our lab. But previous studies, for example by IBM and Google, have also shown that reasonably high inlet temperatures (lower than 40°C) have no significant effect on the reliability of the electronics.
Modern data centers should avoid servers that cannot cope with higher inlet temperature at all cost as the cost savings of free cooling range from significant to enormous. We quote a study done on a real-world data center by Intel:
"67% estimated power savings using the (air) economizer 91% of the time—an estimated annual savings of approximately USD 2.87 million in a 10MW data center"
A simple, solid and very affordable server without frills that allows you to lower the cooling costs is a very good deal.
Facebook
Twitter
RSS
48 Comments
View All Comments
bobbozzo - Tuesday, February 11, 2014 - link
"The main energy gobblers are the CRACs"Actually, the IT equipment (servers & networking) use more power than the cooling equipment.
ref: http://www.electronics-cooling.com/2010/12/energy-...
"The IT equipment usually consumes about 45-55% of the total electricity, and total cooling energy consumption is roughly 30-40% of the total energy use"
Thanks for the article though.
JohanAnandtech - Wednesday, February 12, 2014 - link
That is the whole point, isn't it? IT equipment uses power to be productive, everything else is supporting the IT equipment and thus overhead that you have to minimize. From the facility power, CRACs are the most important power gobblers.bobbozzo - Tuesday, February 11, 2014 - link
So, who is volunteering to work in a datacenter with 35-40C cool aisles and 40-45C hot aisles?Thud2 - Wednesday, February 12, 2014 - link
80,0000, that's sounds like a lot.CharonPDX - Monday, February 17, 2014 - link
See also Intel's long-term research into it, at their New Mexico data center: http://www.intel.com/content/www/us/en/data-center...puffpio - Tuesday, February 18, 2014 - link
On the first page you mention "The "single-tenant" data centers of Facebook, Google, Microsoft and Yahoo that use "free cooling" to its full potential are able to achieve an astonishing PUE of 1.15-1."This article says that Facebook has a achieved a PUE of 1.07 (https://www.facebook.com/note.php?note_id=10150148...
lwatcdr - Thursday, February 20, 2014 - link
So I wonder when Google will build a data center in say North Dakota. Combine the ample wind power with cold and it looks like a perfect place for a green data center.Kranthi Ranadheer - Monday, April 17, 2017 - link
Hi Guys,Does anyone by chance have a recorded data of Temperature and processor's speed in a server room? Or can someone give me the information about the high-end and low-end values measured in any of the server rooms respectively, considering the equation temperature v/s processor's speed?