Since I have been traveling extensively over the last few weeks I have not been able to give much thought to our blog. However, the travels have also provided a little fuel for some comments.
First, I continue to be surprised/pleased to hear more and more presentations and discussions about evaporative cooling of data centers. It seems that "the big guys" get it...cooling data centers costs a fortune using compressors/chillers and the servers can handle much higher temperatures than people realize. If you run down the roster of large international web service or cloud service providers you will find that most of them have already implemented evaporative cooling or they have it in the construction plans.
As great as this is there are still market forces that are conspiring against this highly efficient cooling solution. One is the concern over humidity levels in the data center. This concern is compounded by the common use of relative humidity as the conversation point when it is actually absolute humidity that should be considered. This topic will likely be a point of debate for a long time to come since some of the larger companies have concluded that absolute humidity doesn't matter in their facilities...especially with 2 or 3 year server refresh rates...and other members of this progressive group are not sure and choose the "safe path" of limiting absolute humidity or dewpoint in their spaces.
The one area where it seems that all of the large players agree is with regard to temperature. It is virtually universal that ASHRAE 9.9 recommended guidelines are acceptable and, for many of these users, ASHRAE 9.9 allowable temperatures are OK.
The challenge for the industry is still finding a way to filter this information and confidence down to smaller operators and owners. I have heard it described as an education issue but is that truly the case? It is hard to find a computer or data center related design publication these days that does not promote higher temperatures as a feasible solution for cutting operating costs. Are we just too busy to read these articles or do we not believe the wealth of research and experience that backs up the statements?
At a recent conference on data center design I sat at a lunch table with a group of design engineers and a manager of 13 data centers. When asked how he learned about managing those centers the response was that he was self taught by attending conferences and talking to "experienced" data center managers. So his knowledge of the work by ASHRAE and others was not a major factor in deciding on appropriate operating temperatures. What he was learning was what these other managers had been doing over the last decade...going back to the "old days" where electric costs were low and low data center temperatures were the norm and research had not shown that to be unnecessary.
So, if education is the issue then how do we go about it? What mechanism will get the message through the daily clutter of information and time demands? I don't have the answer...if I did I would implement it immediately....but it seems to be a key to moving the industry forward.
No comments:
Post a Comment