Will Oracle release its own Linux distribution? Jason Perlow says it
As I write this, there’s buzz going around (again) that
Larry Ellison (who looks remarkably similar to Lex Luthor’s
evil, megalomaniacal dad on Smallville) and his legion of robots
are planning to release come out with a Linux disribution,
presumably as the premier platform for running
"i">Oracle 10g. Great! Just what we need —
another Linux distribution.
Now, don’t get me wrong, I’m all for diversity in
the Linux landscape. I think there’s room for a good, say,
half-dozen parent distros and a handful of offspring to fit the
needs of everyone who wants to run Linux. Red Hat,
SuSE Linux Enterprise (and their open source alter-egos,
Fedora, CentOS, and
"i">OpenSUSE), Debian and
Ubuntu (and those derivatives, such as
MEPIS, Linspire, and
"i">Xandros, plus all the cool “Live CD”
utility-things), plus Mandriva should completely — and I mean
completely — fit the bill for the entire community. Throw in
Gentoo for those lunatics that want to build
everything from source, and that’s all the world really
Now, I can understand why Oracle might
think that spinning its own distribution is a good idea, because
let’s face it, a Linux that was pre-loaded or optimized for
Oracle 10g would be nice. Oracle isn’t the easiest software
to install on Linux because it requires a whole bunch of manual
configuration file tweaks to get it to run right, and I think some
better integration with the operating system would be beneficial.
However, I don’t think that further complicating the
landscape and stretching the resources of the Open Source community
even further along company/party/ideological lines is necessarily a
We have way, way too many Linux distros
"i">now. Many of them shouldn’t exist in the first
place. Instead, the goals and novelties of small distriibutions
should probably just get folded into another Linux distribution
family. I know this might be a severe analogy, but we need to look
at the prospect of creating new distros like Planned Parenthood
manages procreation: We really have to think twice before bringing
another one into the world.
People, wear those condoms on and stop forking! Otherwise, the
Linux community might be forced to hand out” safe source” pamphlets
to young and eager programmers attending Linux trade shows and
installfests. Or perhaps we should consider Distro Education
classes in public schools and universities.
Now, I understand why such forking normally occurs. Typically,
the management team of a distro becomes dysfunctional, and then one
or more developers from that distro wanders off the reservation to
pursue something better, such as the case with Debian and
In some cases, such evolution is positive: one distro dies to be
reincarnated anew, holding the birth-to-death ratio constant. And
there are other reasons why making a new distribution or sub-distro
may make sense. Localization can be a big issue, such as with
Chinese, Japanese, Korean, Hebrew, and Arabic, and the various
Indian scripts, where you have to make some major changes in the
code for a local audience to make a Linux distro usable. Politics
and geographical isolation is often a huge barrier for developers
to participate directly in the parent project. Or perhaps
you’re trying to do something really specialized, such as the
One Laptop Per Child project, where you need
to make a special embedded version.
But making a Linux distro just because you want to make your
application more accessible or easier to install or run on Linux is
like birthing another child that the community has to care for.
That’s just stupid.
There are so many ways that a company like Oracle can make the
10g database and its add-on products easier to run on Linux without
having to spin a new distro. Buying one of these Linux companies is
one thing (and my sources tell me this ain’t gonna happen),
but further complicating the landscape is another.
For starters, Oracle should be working directly with companies
like Novell and Red Hat, so that installing the database
isn’t such a pain. Oracle could also engage the Open Source
community, such as the Fedora and OpenSUSE crews, to improve the
Oracle installer and help with the integration, particularly if Red
Hat and Novell are strapped for internal resources. So tuned, it
would be a simple matter for Red Hat and Novell to release SKU
bundles of the operating system and database, even in
specially-mastered DVDs with an integrated install script.
Supporting Oracle widely, on the most popular Linux distributions,
improves the Linux system integrator and reseller ecosystem, and
Please, please Oracle, don’t bring another child in this
world and put the burden on the rest of us.