Editor’s Note: Sun May Rise on Humanity

By Brian Proffitt
Managing Editor

In the early part of 1991, the airwaves in the U.S. were
inundated with a pop tune from C+C Music Factory known as “Things
That Make You Go Hmmm.” “Inundate” is definitely the right term,
because every time I turned on the radio I heard this ditty. Since
at the time I was working as a newspaper editor in western Indiana,
my only radio choices were Pop, Country, and Western–far
from my usual classic rock or indie rock stations. And,
unfortunately, I was in my car a lot.

The early 90s were a dark period in my musical past.

As a result of this needless exposure, I still have this tune
pop into my head from time to time when something out of the
ordinary occurs. Luke Skywalker has John Williams’ symphonic scores
for his soundtrack. I get bad pop music. This week, this particular
piece of my soundtrack got a bit of a workout.

It showed up about the time when the JavaOne keynote from
Jonathan Schwartz revealed an interesting new connection between
Sun Microsystems and Ubuntu Linux’s Canonical, Ltd. It was
interesting for one reason: why was Canonical getting the
Linux-on-Niagara warm and fuzzies and not another distro?

Don’t get me wrong, there is no reason why Ubuntu should not
have gotten the nod and unofficial boost from Sun. They are worthy,
and probably the most deserving. Canonical’s Mark Shuttleworth has
been touting for quite some time that Dapper would be a good fit
for the enterprise and with a clear shot at the newest Sparc
platform, Ubuntu has a great chance to make that happen. (Keep in
mind that this is all speculative, since no formal arrangement was
announced between the two companies–yet.)

Here’s what I don’t get. By showcasing Ubuntu, Sun deliberately
snubbed Red Hat and Novell. Now, I know that for Sun, it’s mostly
about OpenSolaris; and Red Hat Enterprise Linux and SUSE Linux
Enterprise Server are the main Linux competitors for that flavor of
UNIX. So on that level, a snubbing makes sense. Particularly with
the latest reports out there that Niagara is a good deal faster
than Intel’s Itanium chip. If true, then any Linux that’s behind on
porting to Niagara might be at a disadvantage in the 64-bit

Granted, Niagara’s specs are open, so Red Hat and Novell don’t
need Sun’s backing to port to that platform, but there’s a real
benefit to getting the O-fficial Space Ranger Seal O’ Approval from
any hardware vendor, and right now, it looks like Ubuntu is the
only Linux distro getting that seal from Sun.

With all of that in mind, however, I can’t figure out what’s in
it for Sun? Yes, they shake up their main OS competitors, but what
does Ubuntu bring to Sun? Even though Sun may not like Red Hat and
Novell right now, it seems pretty obvious that their rivals are
capable of bringing a lot more server business to Sun than Ubuntu
could. Because if it’s mostly about OpenSolaris, never forget that
it’s all about selling hardware for Sun. As much as I will cheer
Ubuntu on, I don’t see them bringing a lot of existing Linux
customers to Niagara. I would expect that Red Hat and SUSE users
would be more numerous and therefore more of value to Sun.

I see three possibilities why this decision was made:

  • Sun want to get more leverage with Red Hat and
    This is rather obvious, and Occam’s Razor dictates
    that it could be the reason by its sheer simplicity. This
    demonstration of Ubuntu on Niagara may be a way of Sun warning the
    Big Two Linux distros that “we don’t need you.”

    As straightforward as this is, I don’t think this is the case.
    It could be; goodness knows there are enough egos at Sun to think
    in such an arrogant manner. But while I have disagreed with Sun’s
    methodologies in the past, I don’t think they’re stupid.

  • Sun knows something we don’t. The idea that
    Sun perhaps tried to play nice with Novell and Red Hat first and
    got rebuffed is an attractive one. It means Sun kept its eye on the
    ball to sell more servers. But it also may mean that they
    discovered that Red Hat and/or Novell is about to announce a
    hardware deal of their own and left Sun out in the cold.

    I like this idea, in that it is always great to hear that
    hardware vendors are releasing and supporting Linux distributions
    for any customers. But I have some doubts, since Red Hat and Novell
    seem to be doing fine without being too attached to a particular
    hardware platform. And nobody seems to have a burning desire to
    board Itanium, which is fast becoming the Mary Celeste of the
    platform world.

  • Sun is about to acquire Canonical and, therefore,
    This one is the most outlandish, but I think could
    happen. Here’s why.

Way back in December of 2004, a Merrill Lynch analyst made that
rather bold statement that in order to be taken seriously in the
Linux market, Sun needed to acquire Red Hat or Novell. It certainly
fell in line with one Sun executive’s comments made in August of
that same year, indicating that Sun was looking into acquiring
Novell in order to gain a better market position against IBM, which
was clearly allied with Red Hat Linux.

That Sun executive was then-President and COO Jonathan

For whatever reason (I suspect too high of a price tag), Sun
never acquired Novell. But I am left wondering if now-CEO Schwartz
still has his 2004 plan in mind when he cozies up to Canonical.
Because as an acquisition, Canonical can bring a lot more to Sun
than it would as a partner.

Acquiring Canonical would mean that suddenly Sun would have a
full-fledged (possibly) enterprise-ready Linux distribution with a
huge community following. If Sun handled it right, that community’s
energy and resources could expand into other Sun open source
products, such as OpenOffice.org and maybe even OpenSolaris itself.
(Not holding my breath for that one, but getting more
teams on OpenOffice.org would be a big plus on its own.)

Getting Ubuntu gets Sun some street cred in the community, as
well as providing Sun customers with a version of Linux that Sun
can control the direction. Working with Red Hat and Novell is all
well and good, but if they decide to take their products in an
engineering direction that you weren’t expecting, you the hardware
vendor are constrained to going along with them and selling your
customers what the Linux vendors think is important. If you have
your own Linux distribution in house, well, that’s something quite
different isn’t it?

As I mentioned earlier, Ubuntu and Canonical get instant access
to a huge array of enterprise-level customers thanks to the power
of Sun’s marketing and sales channels–if they connect with Sun.
Ubuntu’s good enough to eventually get to these customers on their
own, but in this way they’ll skip all the preliminaries and get
right in the enterprise game.

Back on the Sun side, pulling Ubuntu into the fold will also get
Sun something that IBM doesn’t really have: a way into the
downchannel markets, such as small to medium-sized businesses.
Ubuntu, after all was a very solid desktop OS first, and could be a
good match for any client platforms Sun might want to sell

Is this feasible? Yes. Managing the two technologies would be
fairly simple, because even though Solaris and Linux are different
operating systems with different kernels, much of their software
components are the same. Heck, the Nexenta OpenSolaris distribution
already uses Debian package management… right down to the
Synaptic package management tool found in Ubuntu Linux. Both OSes
are aimed right at the same customers, too.

Is this a viable business option? I think it is. Each company
needs what the other has, and the price tag for buying Canonical is
not going to be as high as buying Novell.

The only question is, if this does happen, will Sun be able to
keep the goodwill of the Ubuntu community going? Or will they end
up buying technology with no developers?

Things, I daresay, that will make us all go hmmmm.