[Swift-devel] pilot job paper

[Mihael Hategan]

Thanks.

I think I tried previously to find the raw data they used, but couldn't
get much out of the paper in that direction. And that seems to be
essential for the numbers I want. One could use QBETS (if it was exposed
as a service) to get the needed data instead of running the actual test,
but I think that would be a weaker argument.

Ultimately one can fall back to the argument that there is some constant
cost in all cases for queuing systems, a cost which is relevant for jobs
of the order of 10s, but it would be more interesting to show that this
is a good strategy for a wider range of jobs.

Mihael

On Thu, 2011-07-21 at 09:19 -0500, Ian Foster wrote:
> E.g., I used it
> here: http://ianfoster.typepad.com/blog/2009/08/whats-fastera-supercomputer-or-ec2.html
> 
> 
> 
> On Jul 21, 2011, at 9:18 AM, Ian Foster wrote:
> 
> > I think that you may find it useful to look at the queue time
> > predictor that Rich Wolski developed. QBETS I think is the name?
> > 
> > 
> > Ian.
> > 
> > On Jul 20, 2011, at 4:41 PM, Ketan Maheshwari wrote:
> > 
> > > 
> > >         
> > >         
> > >         I need queuing time vs. advertised job walltime on various
> > >         clusters
> > >         (with various/random degrees of utilization). That's to
> > >         see whether it's
> > >         useful to have coasters at all.
> > >         
> > >         The number of jobs is an orthogonal dimension (i.e. we may
> > >         want to
> > >         measure the queuing time vs. #of jobs for various
> > >         walltimes, but later).
> > >         The actual job duration is not relevant. The amount of
> > >         data is not
> > >         relevant.
> > >         
> > >         Clouds are an interesting environment, but not for this
> > >         particular
> > >         problem. That's because we need to see how much it takes
> > >         to acquire
> > >         resources, not how fast some job middleware is after we
> > >         got hold of
> > >         those resources.
> > >         
> > > 
> > > Do you have any specific environments in mind for these
> > > experiments? For the requirement of various/random degrees of
> > > utilization, we can use MCS local cluster (10 x 64bit + 3 x 32 bit
> > > machines), Beagle, and Ranger. 
> > > 
> > > Regards,
> > > -- 
> > > Ketan
> > > 
> > > 
> > > _______________________________________________
> > > Swift-devel mailing list
> > > Swift-devel at ci.uchicago.edu
> > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel
> > 
> > 
> > _______________________________________________
> > Swift-devel mailing list
> > Swift-devel at ci.uchicago.edu
> > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel
> 
> _______________________________________________
> Swift-devel mailing list
> Swift-devel at ci.uchicago.edu
> https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel