LinuxWorld:
A Corporate Linux
An
overall look at LinuxWorld made it very clear that everything has
changed. Whether that
change is for the better depends, I suppose, on your perspective.
The show was all about the interest, commitment and influence
of major systems and software vendors to the Linux platform and it
was difficult to stand anywhere on the LinuxWorld show floor without
seeing their big booths, their big show crews in their matching
shirts/hats/logos, and the corporate customers they had invited and
delivered.
Smaller
vendors were still at LinuxWorld, but their presence was rendered
less central and important by the noisy voices, professionally
amplified, of the major vendors.
You had to hunt for them in the back, in the Linux Hatchery
(where the little penguin “chicks” are), or tucked in at the
ends of aisles and in other odd spaces.
On
the one hand, the big vendors provide validation, credibility,
marketing know-how, and technical and marketing resources. They are also much better at getting customers who will buy
in volume to come to a show. (And
you could tell they were trying; both IBM and HP had “customer
day” events to take advantage of the rich potential market for
Linux in the New York City area.)
Smaller vendors can benefit from the fallout, as well as from
the chance to meet and network with potential partners and
customers.
But
the Wild West feel of a Linux show of several years ago is largely
gone. Linux pioneers
are still present, but they’re treated as gurus or tribal elders,
rather than as actual leaders of the day-to-day business.
You can tell that by looking at who’s doing the keynotes.
IBM. SuSE/Novell.
CA.
This
article will try to give you a taste of the proceedings at
LinuxWorld in New York City, January 2004 by dipping into keynote
sessions, meetings, briefings, and press announcements.
IBM’s
Ross Mauri was one of the LinuxWorld keynoters. His speech was also a part of IBM’s customer day. It
started off with the exciting pronouncement that IBM believes that
$25 billion in Linux revenues have occurred, worldwide, in the last
three years.
IBM’s
emphasis is on the freedom and choice that Open Source can provide.
They believe that this freedom, and the ability to build
systems based on components from many sources, leads to a lower TCO
(Total Cost of Ownership). That’s
in dispute, as you’ll see when you read the section of this report
from our meeting with Microsoft’s Martin Taylor.
To
IBM, the important news is that Linux is a mainstream platform now.
They offered proof points from the leading analysts:
•
Forty-five per cent of mid-size businesses are using or experimenting
with Linux. (Gartner)
•
Linux server shipments will grow by a CAGR of more than 28% through 2007.
(IDC)
•
A survey of more than 500 users in North American found Linux was in use
at more than 80% of their sites (72% on servers; 15% on desktops).
(SG Cowen)
Mauri
went on to give examples of Linux in use in all kinds of platforms,
Super Computers to embedded systems, in all types of applications.
Mauri also touched on the newly emerging Linux Desktop
activities, especially the forming of a Desktop group at Open Source
Development Laboratories and IDC’s projections of a much faster
growing Linux desktop market.
NOTE:
In fact, today (2/15/2004), IDC is quoted in Business Week as
noting that Linux Desktop sales in 2003 more than doubled to 3.2%
(more than Apple Macintosh) and will rise to 6% by 2007. We’d guess that if the non-US initiatives are one-quarter
as successful as their sellers are hoping, that number will be
closer to 10%.
Perhaps
the most exciting part of the Linux story is the interest of
governments, both in the US and abroad, in considering Linux as part
of their IT plans. In
some cases, this is all about building out an IT infrastructure
under highly constrained budgets.
In other cases, it’s more about the possibility of
nurturing a native software industry that could become an engine for
export growth. In still
others, it’s all about an interest in security or in the
opportunity to customize systems through open source code.
IBM
wrapped up with a discussion of advances in Linux technology and the
IBM commitment to supporting the refinement and growth of Linux and
Open Source across all of its products and platforms.
HP
HP
has announced that it has generated $75 million in Linux revenue as
a result of its Solaris Migration program that began in early 2003.
HP also announced results of its "Solaris-to-Linux
Lifeline" migration program with 50 new customers in three
months. More
information is available at: www.hp.com/linux.
Novell
In
their keynote, Novell’s Jack Messman declared, “We’re a
billion dollar company wagering our future on Linux.”
He went on to say that Open source is all about thinking
differently – changing control so that vendors have less control
and customer have more.
He
also couldn’t resist a little wave in the direction of Redmond,
commenting that “Fifty to eighty per cent of proper studies show
Linux has a lower TCO,” implying that perhaps some of the studies
being offered by Linux competitors might have a biased point of
view.
What
I liked best were his comments on the way to make money on the Open
Source model. “Novell
had a bad case of NIH (not invented here),” Messman admitted.
But Open Source is all about taking advantage of collective
innovation and then leveraging your paid development staff to add
value on top of that community investment.
More info at www.novell.com.
Sun
Sun
wanted LinuxWorld attendees to know that they are really serious
about Linux. They’ve
moved beyond a Sun distribution – their original idea – and now
base their offerings on standard Linux distributions plus their
hardware and software offerings.
Their Java Desktop runs on Linux and their latest release of
their Java Enterprise Services is also available on Linux.
Of
course, Sun keeps telling us “but you can also run Solaris x86 on
our Intel and Opteron offerings.”
We wonder just who will be buying Linux from Sun?
Only, we suspect, well established Sun customers who just
happen to want some Linux in their product mix without adding
another vendor – until Sun takes that “but” out of their
marketing. More info at
www.sun.com.
Veritas
Veritas
is doing two things these days that Linux folks will be interested
in. It is adding Linux
support to traditional Veritas products.
For example, they have added SuSE support to their Red Hat
support for more international coverage and are also supporting
VMWare. At the same
time, they are building a suite of management applications (our
Virtual Computing portfolio), with an emphasis on supporting Utility
Computing. More info at
www.veritas.com.
Penguin Computing
We
met old friend Enrico Pescatori at Penguin, a 6-year old company
focused on selling high-end Linux systems with an emphasis on
clustering, scalability, 64-bit computing, and the ability to
support numerically intense applications.
They company sells mainly servers and is the developer (via
acquisition) of the commercial version of the Beowulf clustering
software for Linux, Sclyd. More
info at www.penguincomputing.com.
Computer Associates
CA’s
Sam Greenblatt likes to say “We are in the Linux generation.”
In his keynote at LinuxWorld he took users on a tour of where
Linux is today – especially all the places you might find Linux
implementations, both servers and desktops, and a glimpse into Linux
tomorrow. Particularly
fascinating were all the places Sam was able to point to Linux
running invisibly from Sharp PDA’s to Cisco switches to Tivos.
More info at www.ca.com.
NetIntegration
(Nitix)
I
had never heard of tiny Nitix before I saw them at LinuxWorld.
That’s too bad because they have a good idea – one that I
suspect is required for Linux to “catch on” in the SMB market.
Nitix markets what they call autonomic Linux which turns out
to be a server, pre-loaded with Linux and a set of core IT
applications: File and
Print, Internet Server, Routing, Firewall and so forth, but also
including messaging, collaboration, and database.
They bill their box (I’d call it an appliance) as self
healing, self-diagnosing, and sell-installing and target it at firms
with 5 to 250 end users.
The
magic is in a special chip that works with their software.
They’ve sold 7,000 boxes, mainly in the US, via resellers.
Now they’re prepared to also sell the software plus the
chip to resellers who want to put it on their own box, perhaps with
additional software, going to vertical markets.
We think it’s an intriguing idea.
More info at www.net-itech.com/.
LinuxCare
At
one time, LinuxCare was one of those dot.com age high-flyers,
destined to make its founders rich.
It didn’t turn out that way – and neither did its
original idea of providing support to Linux users in big companies
because the big systems vendors thought they’d like that business
themselves.
But
LinuxCare has had a second act, offering software for rapid
provisioning – in less than two minutes.
They can rollback to previous states, do uninstalls, and
because they’re highly compressed they can essentially (depending
on storage capacity) save as much of the system’s previous history
as desired. More info
at www.linuxcare.com.
Microsoft
Microsoft
has been maintaining a presence at LinuxWorld for a year or so now.
Why, you might ask? We’d
guess it’s partly a listening post, partly for positive marketing
(buy us, not that) and partly for negative marketing (you don’t
want to do that). The
booths keep getting bigger and they keep adding Linux Outreach
programs so they’ll have something to talk about.
We
met with their newest exec for this sector, Martin Taylor.
Like all Microsoft execs we’ve met, he doesn’t really
believe that Linux is a real competitor.
He wants to bound it off in a corner somewhere.
It reminds me of the days when I’d go to a big systems
vendor’s briefing and they’d tout their UNIX offering and bound
Windows/NT off in a little corner. Microsoft (and its customers), of course, never paid any
attention. Eventually, the systems vendors noticed that Windows was
a really big market and they climbed on board.
After all, customers, not vendors, determine the eventual
outcome here.
Microsoft’s
newest pitch (backed up, of course, with an analyst study) is that
Linux is too expensive because the cost of interoperability is too
high. In case you
can’t figure out what that means, let me provide translation
services. They mean
that a huge number of applications are written to the Microsoft
Windows APIs (estimated number 50,000+ and we’re only talking
commercially available ones, not all the in-house stuff).
They are the big barrier to users who’d like to move to
Linux; they can’t just move their applications with them.
There are some schemes for moving software or living in
hybrid worlds, but none of them are cost and/or effort free.
Software
development is driven by two issues:
the size of a market (the bigger the market – or, rather,
the bigger developers think it will be – the more software gets
written) and its maturity. Linux
is still too small and too young to have attracted an equivalent
amount of software.
If
you’re passionate about moving to Linux you’ll say, “Yes,
but…” and you’ll
mean, yes, but, I don’t need 40,000 packages, I just need the four
(or the 14 or the 44) that I use. You’re right.
The trouble is you don’t know which ones you’ll need next
month or whether they’ll be there.
If you’re a developer you won’t spend much time worrying
about this; you could always write a missing piece.
If you’re a customer, it can be a major issue.
On
the other hand, time and market dynamics can move mountains.
For one small example, I noted with interest today that the
Lindows.com Click Once Software Warehouse has gone from 1,100
software packages last year to 1,830.
Most of the infrastructure software for servers is available
on Linux now. Software
availability to guarantee interoperability is a problem, but it’s
probably a transitory problem that is getting better over time.
Also,
we’re changing the architecture.
As we abstract how we connect things up, becoming more and
more dependent on Web Application Servers and Web Services,
operating systems and operating system dependent APIs become less of
the issue. Open
Standards that cross operating systems environments become more
important. So perhaps
this is an issue today, but not an issue at all in n year.
We’ll take bets on where n is. I’d guess three to five
years for mainstream applications.
Anything up to infinity for some obscure things.
But there are always XML wrappers.
For
more on Microsoft’s point of view www.microsoft.com.
More
information on Linux, I recommend LinuxWorld Magazine, http://www.linuxworld.com/.
(back to top)
