[Rocks-Discuss]Re: Cluster Poll Results (tangent into OS choices)
Philip Papadopoulos
phil at sdsc.edu
Tue Nov 4 16:05:21 EST 2003
Daniel Widyono wrote:
>I just asked Mason about the source of his statement, and he referred me to
>RedHat's own site, from which I found at
>
>http://www.redhat.com/about/corporate/trademark/guidelines/page9.html
>
>the following quote (taken out of the Educational Institutions paragraph):
>
>"This permission is not applicable to Red Hat. Enterprise Linux. or any Red
>Hat subscription product. Of course, you are always permitted to
>redistribute the code without utilizing Red Hat's trademark so long as you
>otherwise comply with the GNU General Public License and Red Hat's Trademark
>Guidelines."
>
If read the _entire_ paragraph, this refers to a redistribution of their
binaries (which we did for distributions <= 7.3). For IA-64, we
recompiled everything that was on RedHat's open-source
_source_ directory. I very much doubt that the enterprise-specific
crown jewels (failover, etc)
are in this source directory. In particular, we are not redistributing
redhat-created binaries. We went
down the path of trying to get permission to re-distribute IA-64 to a
small select group of
folks, but could not and therefore we didn't -- Hence, we "figured" out
how to rebuild a
complete IA-64 distro from the open-source sources.
Also, we are only working with the advanced workstation (AW) source
tree. I don't know (and don't
care) if this the same base set that is used in the advanced server (AS)
and enterprise server (ES)
version that Redhat sells as well. Our _current_ clustering needs are
met quite well the re-compiled
open-source rpms of AW. And we are not trying to re-engineer RedHat's
entire line (even though
some may think we are). If one needs AS or ES, the web address is
http://www.redhat.com.
So, are there things in AW that aren't in Rocks. Probably. I haven't
looked in detail.
Are there things in Rocks that aren't in AW. Absolutely (go check our
cvs repository).
Are there open-source add-ons in Rocks that we didn't author? Uh. Yeah.
Try a whole
litany of base cluster tools. MPICH, GM, SGE, Globus, Condor-G. ... .
Are there open-source
things that Redhat didn't author? Duh.
So if it's all open source, what does redhat sell? Services, patches,
updates, notifications, integration.
All of this is very very valuable. They make key contributions to linux
development and have authored
a whole bunch a critical software -- Their (open-source) packaging
format is de-facto standard. Even
SuSE uses it.
We like Redhat (alot), want to support them (both morally and with $$).
There is, however,
a reality of how much money people have and how much they are willing
to spend. RedHat will find
that balance (I hope) for clusters, universities, and others. I believe
that most folks agree
that O($200/node/yr) does not match either the amount of money people
have or
how much money they are willing to spend for the support in a clustered
environment.
-P
>
>cAos's work is sounding mighty tasty. Thanks for the cross-post, Mike.
>
>Dan W.
>
>
>
>>Also, nice to see you cross posted to the rocks-discussion, for the
>>benefit of those on the beowulf list, Mason Katz (mjk at sdsc.edu) had an
>>informative reply:
>>https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/2003-November/003567.html
>>
>>It would appear as though Rocks is free and clear to openly redistribute
>>RHEL SRPM-rebuilds; this is an interesting loop-hole:
>>
>> - Rocks released by an academic institution, which means it has a
>>license to use the RedHat trademark. This also means no one can charge
>>for Rocks software (only support).
>>
>>
--
== Philip Papadopoulos, Ph.D.
== Program Director for San Diego Supercomputer Center
== Grid and Cluster Computing 9500 Gilman Drive
== Ph: (858) 822-3628 University of California, San Diego
== FAX: (858) 822-5407 La Jolla, CA 92093-0505
_______________________________________________
Beowulf mailing list, Beowulf at beowulf.org
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
More information about the Beowulf
mailing list