[Beowulf] Servers Too Hot? Intel Recommends a Luxurious Oil Bath

Robert G. Brown rgb at phy.duke.edu
Wed Sep 5 09:15:44 PDT 2012


On Wed, 5 Sep 2012, Lux, Jim (337C) wrote:

> But as long as we're talking quarrys and such, what about the scheme of
> building a big pit to fill full of ice during the winter, and melting it
> during the summer. (assuming you are in a
> less-than-wonderful-un-California-like climate where this would work.)

Or, just place all of your cluster in Antarctica.  Or somewhere on the
Greenland ice pack.  Or in Tibet.  There in any mountainous region where
the cluster is up there 4 or 5 km at a mean air temperature that is
"cold" so you don't need refrigeration, only circulation of outside
ambient air.  If you want to get all fancy, build your nodes so that
they are weatherproof and sit underneath several square meters of solar
panel and battery backing so that they run "for free" if you build them
in a cold, high, desert like the Andes.

The problem with all of these schemes is that they all cost too much
compared to just buying a pod, lining it with over the counter shelving
equipped with wiring harnesses back to a central power supply, sticking
a few tons of AC on the ends mixed up with fans, and plugging the whole
thing into a 440V grid.  Need more capacity, add a pod.

I recall that this almost perfectly describes Google:

   http://en.wikipedia.org/wiki/Google_Modular_Data_Center

except that they use larger shipping containers, not smaller ones, at
least sometimes.  They are pretty close to the chest with their actual
technology.  They solicited me a short while ago to come join their
cluster group (which I could not do, being enormously overcommitted
already with my own entrepreneurial gig and more) and I kinda wish I
could have pursued it as it would be fun and profitable and if I had I'd
know a lot more about how they rig things.

But of course if I had done this I would have had to sign and NDA and
hand over my firstborn son (or kill you) if I told you...;-)

    rgb

>
>
> On 9/4/12 11:56 AM, "Douglas Eadline" <deadline at eadline.org> wrote:
>
>>
>> Of course those massive Zetta scale systems will live in huge
>> multi-story oil tanks that have been placed in old quarrys,
>> which provide bedrock support for the tank and a geothermal
>> heat sink. The sysadmins must operate oil swimming robots and even don
>> oil scuba suits to service hardware. Don't forget the Sterling engines
>> using the heat from oil to return a bit of useful work to the system.
>>
>> --
>> Doug
>
> _______________________________________________
> Beowulf mailing list, Beowulf at beowulf.org sponsored by Penguin Computing
> To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
>

Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu




More information about the Beowulf mailing list