[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] What is the industry standard for grid computing?

On 4/27/10 5:10 PM, Lee Mitchell wrote:
> I realize the persons on this list are not necessarily unbiased, but I
> would like to ask:
> What would you say is(are) the industry standard(s) for compute grid
> software?   By standard, I mean most commonly adopted, not a
> formalized standard.  I believe Condor should be on such a standard
> list, but what else has wide enough adoption to rank as an industry
> standard?
> Does anyone know of sources with numbers on proportions of "market
> share".  I use the term loosely because I know a lot of the grid
> "market" is at publicly funded institutions.  But I'm also interested
> in such numbers for corporate grids.


That is a super loaded question.  I use Condor heavily and think it is
great, but I consider it a real stretch to call it grid software.  It
has several grid-like capabilities, and it can be integrated into a grid
computing environment, but it is still predominantly a job management
system for scheduling queued computations to a pool (or cluster) of
computers.  That said, Condor is more "grid like" than SGE, which has
the word "grid" in its name, especially with its built-in integration
with Condor-G.

It turns out that the features of grid computing -- heterogeneity,
distribution, federation, dynamism, data management, standardized
interfaces, etc -- relate to corporate computing issues mostly around
areas of resource management, and for various reasons this is in such a
way that the independently developed virtualization technologies of the
last decade (building, of course, on virtualization concepts that go
back to VMS and probably further) actually address most of the corporate
requirements in a way that means "cloud" computing (in whatever form:
IaaS, PaaS, SaaS) is more interesting and applicable than "grid" computing.

The Open Grid Forum is the best place to find "standards" around grid
computing.  The Globus Alliance and Globus Toolkit provide the
foundational components of most grid infrastructures, although the
European EGEE project is catching up (or has perhaps surpassed Globus)
in implementing standards-based grid software components as part of the
EGEE/WLCG/gLite infrastructure.  Most large scale grid infrastructures
(and in fact all that I can think of) rely on Globus in some way.  EGEE
tries harder than most to be a standards-based grid.

I was part of a European Union funded task force to evaluate the
technologies and standards.  If you do a Google search on "ETSI Grid"
you should find the reports we produced.  Unfortunately these are going
to read like an EU-funded report, so you may do better looking for an
industry-oriented white paper from some consulting company who has
less-than-average vested interest in a particular technology or position.

Feel free to contact me directly if you want to speak further.


fn:Ian Stokes-Rees, PhD
org:Harvard Medical School;Biological Chemistry and Molecular Pharmacology
adr:250 Longwood Ave;;SGM-105;Boston;MA;02115;USA
title:Research Associate, Sliz Lab
tel;work:+1.617.432.5608 x75