[gridengine users] Beware Univa FUD
rayson at scalablelogic.com
Tue Nov 15 06:57:34 UTC 2011
Killing or bankrupting closed source companies is usually not the goal
of open source projects. We are just trying to rebuild the open source
Grid Engine project that was founded by Sun Microsystems in 2001 and
ended by Oracle in 2010 (don't blame Oracle, and see below to find out
For almost 10 years (and more than 10 if you count the free
Linux+Solaris SGE binaries with unlimited license before the birth of
the open source project), the market share of SGE was always growing
because it was free and we had many who contributed by answering
questions on the list (many said on the original "users" list that
they got faster & better answers than what they could get from Sun's
paid support). This created the ecosystem that includes system
integrators, users, and other open-source projects & 3rd-party
companies that offer services on top of SGE.
As SGE is the enabling technology, we just can't let it become closed
source or open core, as those are really "non free" according to
today's open source standards. As chosen by Oracle as the follow-on
project, I have the obligation to make sure that open source Grid
Engine is free and open source, and not commercial or open-core. I
believe Dave is trying to do the same with his fork as well.
I am not interested in contributing to open core projects because I am
not interested in making contributions to other companies' bottom
line. However, I am also not interested in destroying other companies.
As I mentioned on the list, I know OLAP cube fundamentals & data
warehousing (+ I know a bit about Oracle Exadata & Netezza), and I
know people who work as ETL consultants. I believe they can quickly &
properly build one for Grid Engine accounting & reporting. If we were
to hire one of those consultants (contract only, no $$ for full-time
yet!), I believe we could get something like UniSight in less than 6
months, and can directly compete with Univa's offering. However, this
is not something we plan to do, because 1) we should allow people to
build useful & value-added products on top of open source Grid Engine,
and less importantly: 2) the front-end for most OLAP cube solutions
are not open source, and thus we will end up with a closed source
Don't blame Oracle, it has a different market segment and open-source
SGE may not fit Ellison's vision and thus it needed to end SGE as an
open source project. Oracle could have just shutdown the project
without saying a word. Instead, Oracle not only picked a follow-on
project (and in fact also recognizes Dave's project as well), it also
offered the project data & materials to us & Dave (in the end, I did
not have a chance to download them because I was busy with other
things when Dan offered me those files, and when I was less busy Dan
has already left Oracle). Since Oracle's announcement of ending open
source GE, almost every single batch system vendor has some sort of
Grid Engine migration program and/or uses FUD against Grid Engine, and
the only one that is not using those tactics is Oracle.
Finally, we have some basic understanding with Oracle that if Oracle
wants to have the open source Grid Engine maintainership back, it is
always welcome to do so (and with our code changes too). And if I
recall correctly, Dan also mentioned this in this RCE Cast:
With this in mind, I doubt Oracle would suit Dave, Ron Chen (who
founded OGS & uploaded most of the code) or myself (OGS maintainer),
as 1) the money Oracle gets in the end would not cover the invoice it
gets from the lawyers, and 2) why it needs to go through all the
trouble to get what it can take back just by asking us, and 3) Oracle
recognizes the OGS & SoGE projects, so I doubt Oracle could get a lot
of $$ by claiming for "damages".
I guess, one can say that Oracle might want to kill all the open
source batch systems, and it could do so by shuting down us all.
However, there are just so many free & non-free batch systems, all
with similar features (or at least good enough features). I don't
think ending 2 open source batch systems could bring lots more revenue
for Oracle Grid Engine. And one of the biggest competitors is earlier
versions of SGE - I know many sites that are running earlier version
of SGE (some with 6.0uX) happily.
In the IP front, I think we have more to worry about, but Oracle is
really not the biggest threat to open source Grid Engines (and I hope
we could end using Oracle to FUD open-source Grid Engines).
2011/11/12 Chi Chan <chichan.gridengine at yahoo.com.tw>:
> If Univa thinks that the reputation is not important, we should give what Univa deserves.
> How about maintaining a web site to show people all the FUDs that Univa marketing sales are using, and tell people to be aware of this company.
> In the early days of Linux vs. Microsoft, MS used to do its own studies to show how bad Linux is. And Linux defenders maintain web sites to show why MS cannot be trusted by listing all the misleading facts that MS uses against Linux.
> Also, we can let other mailing lists know (in fact anyone who is involve in HPC know), what kind of company Univa is, so that Univa gets no customers, and then it can finally exit and shutdown.
> ----- 原始信件 ----
> 寄件者： Dave Love <d.love at liverpool.ac.uk>
> I've no idea why they feel the need to attack underhandedly --
> specifically amongst colleagues and potential employers -- the
> professional reputation of someone who mis-guidedly offered support, but
> others should beware.
> [Apologies not to have been able to respond to things here recently, and
> there seems to be plenty that needs a response if I can find time.
> Anyhow, now for something more constructive.]
> users mailing list
> users at gridengine.org
> users mailing list
> users at gridengine.org
More information about the users