www.internetnews.com/ent-news/article.php/2243431
|
By Erin Joyce August 1, 2003 Q: When's Mad Hatter [Sun's Microsoft-compatible, Linux-based desktop package of applications] coming out and in which industries will it be focused? Probably September or October. For industries, typically large, homogenous workforces such as retail outlets, big branches. We're testing about 4,000 units right now. We have three principal customers: developers who build things, CIOs who pay for things, and operators who then have to deal with the aftermath of applications that were built and bought and make them all work together. When it comes to products, they range from servers and storage that are behind the network, to the devices that present the network, to the devices that authenticate to the network. Those are the three tiers that we see in the user space of how software is built. There's an incredible diversity coming out today on the desktop. Everything you see today that is on the Internet will have a microprocessor smart card on it. Look at the Best Buy [fake Web site] fraud. We are coming to a point where strong authentication is just going to be a mandate. You don't make crank cell phone calls for a reason. It's because the carrier knows you. But if the desktop is authenticated, people will create mischief. Therefore strong authentication is going to be a critical element of all network protocols.
Java defined and invented the notion of network API
On the SCO-IBM Linux dispute, and open source The one thing they didn't realize they wanted was the capacity of their
vendors to indemnify them against the risks that went into the products
that
they were using. And just as there are a bunch of mothers of 14-year-olds
getting subpoenas because their kids are sharing music on the Internet,
there will be a bunch of CIOs to testify with respect to the intellectual
property that they are using.
The reason we're familiar with this is because we own our own
intellectual property (IP). And I want to make sure we're really clear
about
our position on the SCO
lawsuit against IBM [over whether UNIX code was misappropriated for
Linux], as well as IP litigation in general.
Open source to me is an irrelevance in the sense that customers don't
want products based on whether they are open source or not. They buy them
based on whether they are better quality, are faster, are innovative and
help solve problems.
So why did Linux succeed on Wall Street in Round 1? I would claim
because
it just ran faster on a small Intel box and it ran with potentially higher
quality for the tasks that had been assigned to it, such as running a
batch.
But there are two problems [with open source development]. One is that
IBM appears to have committed a landmark mistake in the [alleged] leakage
of
its IP license from SCO into the mainstream distribution of Linux. My bet
is
that as a result, there's going to be a bunch of end users, who, just like
the mothers of 14-year-olds who trade files, will be getting letters
telling
them that they have an obligation to compensate for the liberties they took
with that IP.
Again, I want to be really clear. This is not a condemnation in any way
of open source. Sun Microsystems is the single largest contributor of open
source technologies on the planet. We are a huge believer of open source.
But I probably have as many lawyers working on my open source products as I
do any other because we also believe that when I go to my customer, I have
to indemnify them against the risks of the IP that I'm selling.
And any vendor that says he can't do that is lying to you. So I find it
interesting that IBM says there's no problem and tells its customers not to
worry but refuses to indemnify them against the risks of the SCO suit.
We're a huge believer in open source, but we're also a huge believer in
IP ownership. I think when you confuse those two you run into problems.
So will MySQL topple the database companies? If their products are
better, absolutely. Does it have anything to do with open source? I don't
think so. What matters is: is it better?
Is Linux better at running a one-way Intel server running Apache
Q: You say you can indemnify the product because you own the IP;
there's various parts to the stack. What of enterprises running, say, JBoss
for application servers? They run it because it's cheaper.
That's an illusion. Did you know that Red Hat has no more free Red Hat?
Q: Yes, but [JBoss] doesn't cost $85,000 for an application server
[license].
Solaris is less than the leading Linux distributor out there. How much
is
JBoss? Well it's free the day you get it. How much is Sun's Linux server?
Free.
Q: But you have to buy [Sun's bundled Orion package with] hardware to
get it.
Absolutely. Software opens markets.
Q: People buy open source products because they are cheaper.
But they're not cheaper. Who said they're cheaper? Some are more
expensive. But again, open source is an irrelevance. Take it off the table.
It's like saying let's go look at the licensing terms under which you get a
product. How much is J2EE Q: The SCO dispute against IBM is seen as an anomaly, a one-time
deal,
even a cash transaction issue. There's nothing to prevent Linux from
long-term showing up on the enterprise high-end the same way Solaris
does.
The problem with that logic is like saying if wishes were horses,
beggars
would ride. If Microsoft owns all the software, what's your play Sun? If
Intel owns all the hardware, what's your play Sun?
If Linux takes over the whole planet, if it scales to be five thousand
ways, well I'd question a couple of things.
Linux is no longer a thing. One company produces Linux dominatingly: Red
Hat. Go poll every customer on Wall Street you will not find anything
beyond
Red Hat (and SuSe).
Q: IBM is saying they'll keep working on Linux. It will scale, it
will
be in these large data center environments. So it's not going to be limited
to Apache Go look at all the Linux running on mainframes that's getting pulled off
because it was a stupid waste of time. Look at all these fanciful claims of
how Linux will run on mainframes. Well, now it's all getting pulled off.
Why? Because it's done.
People qualify to distributions. What was the distribution IBM was going
to run on its mainframe? IBM's distribution. Guess how many ISVs are going
to write to IBM's distribution? Zero.
Whose Linux is it going to be? IBM says it's XYZ. If that's true that
means Red Hat wins the market, which means IBM has a problem because
they've
given the whole operating system to Red Hat. So I think IBM saying Linux
will take over the world is equivalent on their part of saying Red Hat
takes
over the world.
Q: From IBM's point of view it's a global services game though.
That's
why they bought PwC.
So Red Hat and Microsoft everywhere? And Jboss will eliminate WebSphere?
MySQL will eliminate DB2?
Q: IBM doesn't really care too much because they're playing a
services
game.
Well services are valued at one times revenue. And it's a declining
industry. Right now my costs of hiring talent in Beijing are a 10th of what
they are in the U.S., a fifth of what it is in Bangalore.
The only difference between our strategy and IBM's strategy is we
believe
in the consolidation in outsourcing and services with technology. And IBM
believes in the consolidation of outsourcing of services to teeming hordes
of high priced consultants. We believe our strategy in the long run wins.
Q: Which Linux distribution version will you be using for Mad Hatter
[Sun's desktop bundle, which includes MS Exchange, Sun's StarOffice
application suite, Mozilla browser, Java 2, Gnome 2 and Linux]?
Our own. We will likely work with other companies to build it.
Q: Wasn't the Sun version of Linux distribution being phased
out?
Server side absolutely. The thing you really need to think about with a
Linux distribution is what does an ISV qualify to? On a server side they
generally qualify to a very low level set of APIs So the distribution we deliver may in fact have someone else's kernel
inside, because we're not going to build a lot of value into that kernel.
It
will have a huge amount of infrastructure around Gnome, around Evolution,
around Mozilla, around Java, Game, Open Office and Star Office built around
it.
Q: Where do you see [open source's] sweet spot?
Quality matters more than the licensing agreement used to build
something. Typically open source development is done on products that are
very stable and homogenous. Apache is a great example.
Do I believe that Linux will scale? Absolutely. No question. Do I
believe
that Solaris will scale down? Absolutely. No question in my mind.
Ask IT directors. [They say] more source code is not a value
proposition.
'I've got enough source code, thank you very much.'
I have customers every day saying I'd love to give you this system for
you to manage for me. Open source to CIOs is an irrelevance to them. What
they want are cheaper products that run more effectively at higher quality
at greater scale with a lower cost of maintenance over time. That's what
we're going to focus on.
|
| Go to page: Prev 1 2 |