Re: Is anyone running NCL on Amazon's EC2 or other cloud computing environment?

This is streaming off-topic, and for that I apologize. However, as
cloud computing catches more hype, I feel a skeptic ought to show his
head...

One of the key problems (and as yet neither resolved, nor
well-understood by most users of cloud services... nor cloud providers)
is how to determine limits for memory, bandwidth, etc.

Realizing that a lot of cloud instances expect or require virtualization
to run, you have several limitations auto-imposed, some of which you
might be able to control, but some which reside in, e.g., Amazon's
domain. What're the virtual machine limits on memory? What are the
host-hardware's memory, clock-speed and bandwidth limitations? How much
of a physical host do you get when you load that VM up? Some of these
are still imponderables, making me wonder if the potential for cloud
computing has been well-considered.

The idea of a cloud rental, per the wikipedia definition, is very
attractive. How nice it'd be to let someone else manage the commodities
of hpc/htc instead of doing that myself. But for HPC, certainly,
maintaining a well-coupled machine interconnect is a great benefit to
model performance, and for my HTC requirements, maintaining my own
control of storage assets, and storage-interconnect assets, aids in
performance.

gerry

Bridget Thrasher wrote:
> A "cloud" is basically a virtual cluster on which you can rent
> space/resources as needed. Technically, Amazon's EC2 is more for what's
> known as "utility computing," which is what I'm hoping to do. Heavy
> computation on the fly:
>
> *"Utility computing* is the packaging of computing resources
> <http://en.wikipedia.org/wiki/Computational_resource>, such as
> computation and storage, as a metered service similar to a traditional
> public utility <http://en.wikipedia.org/wiki/Public_utility> (such as
> electricity <http://en.wikipedia.org/wiki/Electricity>, water
> <http://en.wikipedia.org/wiki/Water>, natural gas
> <http://en.wikipedia.org/wiki/Natural_gas>, or telephone network
> <http://en.wikipedia.org/wiki/Telephone_network>). This system has the
> advantage of a low or no initial cost to acquire hardware; instead,
> computational resources
> <http://en.wikipedia.org/wiki/Computational_resource> are essentially
> rented. Customers with very large computations or a sudden peak in
> demand can also avoid the delays that would result from physically
> acquiring and assembling a large number of computers."
>
>
>
> On Fri, Dec 4, 2009 at 12:36 PM, Dennis Shea <shea@ucar.edu
> <mailto:shea@ucar.edu>> wrote:
>
> I am not sure what "the cloud" is.
>
> What limits do you have on your current system
>
> %> limit
>
> Look at "memoryuse"
>
> If you need 8GB of memory and you have 4GB not much you can do.
>
> If you can fit the data into memory BUT you are near the memory limit
> and you start doing computations, then the software [any siftware]
> will need/use additional memory.
>
> Good luck
>
> Bridget Thrasher wrote:
>
> I keep hitting memory allocation limits and am thinking of
> moving my computing to the cloud. Has anyone done this successfully?
>
> -Bridget
>
> --
> Bridget Thrasher, PhD
> Postdoctoral Researcher
> Climate Central
> www.climatecentral.org <http://www.climatecentral.org>
> <http://www.climatecentral.org>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
>
>
> --
> ======================================================
> Dennis J. Shea tel: 303-497-1361 |
> P.O. Box 3000 fax: 303-497-1333 |
> Climate Analysis Section |
> Climate & Global Dynamics Div. |
> National Center for Atmospheric Research |
> Boulder, CO 80307 |
> USA email: shea 'at' ucar.edu <http://ucar.edu> |
> ======================================================
>
>
>
>
> --
> Bridget Thrasher, PhD
> Postdoctoral Researcher
> Climate Central
> www.climatecentral.org <http://www.climatecentral.org>
>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> ncl-talk mailing list
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Mon Dec 7 06:32:46 2009