Sun Software Quarterly Town Hall (03/03)

Sun Software Quarterly Town Hall

March 28, 2003

M. Petry Good morning, everybody, and thank you for joining us for the third quarterly Software Town Hall. I'd like to say good morning to everybody here in the Palace Hotel in San Francisco. Thanks for joining us this early. For the folks on the phone, especially our colleagues in Europe, thank you for joining us so late. I hope the next hour and a half to two hours will be beneficial to you.

With us today we have several folks from the software leadership team, including Jonathan Schwartz, who is the Executive Vice President of Software for Sun. Jonathan will, for the next 30 minutes or so, walk us through an update on software, give some priorities, market direction, and expand on Project Orion and such.

Following the short presentation by Jonathan, we will do an hour and a half of questions with members of his leadership team who are here with us in San Francisco today.

We have Alan Brenner, who is Vice President of the Consumer Mobile Systems Group. If everyone here in San Francisco can stop by and check Alan's phones out, he's got some of the yet-to-be-released Nokia phones.

We also have John Loiacono, who is the Vice President of Operating Systems Platform for Sun Microsystems. John oversees the strategy and technologies around Solaris and Linux.

We also have Mark Bauhaus, who is the Vice President of Sun ONE Java Web Services, and his technologies and products include Application Server, Integration Server, the recently announced Web Services Developer Platform Pack, and a bunch of other things as well.

A gentleman who is a bit late today is Juan Dewar, and he'll be joining us shortly. He is in charge of Marketing and Strategy for the Mobile Group, and he works with Alan on the marketing side.

So with that, I would like to transfer it over to Jonathan Schwartz. Thank you.

J. Schwartz Thank you, May. Good morning, everybody, and echoing what May said, if you're in Europe and you're taking the time to call in, we really appreciate it.

What I'd like to do is, in some way, just give you a brief recap from the last time we met and the most significant opportunity we had to clarify and articulate where we were going was our Worldwide Analysts Conference, which was a few weeks ago. During that presentation we really tried to categorize and formalize the plans around some of our core constituencies.

In looking at those - and you're going to hear me continually come back and focus on these, folks, because, especially in the software industry, software is kind of everywhere, and it's very tough to answer a question about what's your strategy with X without really dipping down to look at - and I'm on the first slide now - our three core constituencies.

The first is the developer, and for the most part these are the canaries in the coal mine. These are the folks who are creating the content that makes the world's devices more interesting and more useful, and these are both corporate developers as well as independent software developers. These are developers who write games that run on consumer and mobile platforms, all the way up to the military running signal-processing applications on high-end systems.

So for the most part, the priority of the developer is "show me the money", to the extent that they're an ISV, and really "show me the volume" if you are a corporate developer because you don't want to write something that then locks you into one platform produced by one vendor.

Secondly is the CIO, who, in some sense, although they have the final say over what gets deployed in production, they don't often have a whole lot of control about what the developers are looking at. Their focus right now is TCO, where TCO stands for take costs out, and nothing else. They are ruthlessly focused on acquisition expense, ownership expenses, and … to some very expensive consultants who are no longer that well utilized anymore.

And although the developers really create the momentum around new platforms and new opportunities, at the end of the day, the new CIO has to make the decisions about what goes into the infrastructure; what goes into production. How does this really form the part of my infrastructure?

Lastly in the food chain is the data center operator; lastly because they have the least control over the design and the form of the infrastructure that's built. They basically just get passed new applications, new infrastructure, and told, "Make it work." And more often than not, as I think we've seen on Wall Street - a decade of gorging on new IT - they get given the task of consolidating and rationalizing and trying to make all this stuff work together, which is why, in this particular picture, the data center operator shows up with a helmet and a rubber jacket.

So they are really focused on simplicity and efficiency, and again, the food chain here, just so everyone understands - Sun focuses on developers because developers create the momentum that then creates an opportunity for Sun to deliver the infrastructure to CIO's, who then have the task of making sure that it not only operates efficiently, but can work within their existing enterprise - a challenge they give to their data center operators. So that's kind of the food chain, and it's a bit of a virtuous cycle because, to the extent that platforms are successful, CIO's then seek more of them, that creates more opportunity for developers, and on and on.

So again, when you hear me talk about our strategy, or any of the folks that work for me talk about our strategy, it's dominantly looking at each of these different constituencies and trying to understand the strategies with respect to each. So let's move to the slide on the developer.

Again, for the most part, developers care about volume. I think the news that we have here is better now than it has ever been, and I'll give you a very specific example. I was at our Worldwide Analysts Conference, got up on stage, was talking about our strategies, where we're going, came to the developer section, and said, "I want to talk for a moment about the mobile handsets," little consumer devices, and I've got a few examples up here. Sorry; for those of you who are on the phone; you can't see this.

This is a next-generation phone. It was produced by Nokia. It was just given to Alan and me by the president of Nokia. He said, "Here, here's my phone." These things cost probably about $50. This looks like a regular phone until you realize that when you peel off the numberplate here you end up with a full QWERTY keyboard and the capacity to type with your thumb.

It seems like an interesting little thing - I'll pass it around. It's an interesting way of thinking about an assertion we made, and I think Alan and his team contemplated long before anyone else took it seriously, which is at some point those are going to become a part of not only a more interesting infrastructure that exists for consumers, because it turns out those are phenomenal platforms for monotizing Internet content.

When you're paying eight pounds a download in London for games that come down over it, you can make a lot of money really, really quickly, which is exactly what Vodafone is doing, but at some point they begin intersecting with enterprise infrastructure because the last count we have on this - and I'll get back to my story in a second - 65 million people holding these phones or 100 million people holding these phones also have jobs; they go to work, and obviously they want to check their mail, look at their calendar, maybe check the sales force automation or sales force automation application.

So nonetheless, I was on stage talking about how wonderful the growth was around these different opportunities - and I claimed that there are 65 million devices in the world that are actually Java-enabled devices, so everything from the RIM device to these Nokia phones, Motorola phones, and I was taken to task after that session by some folks who were more knowledgeable about the audited numbers, and they said, "Actually, the number, you blew it." I thought, "I got the number wrong. Marketing gave me numbers that were too aggressive. Maybe it's only 50 million." Their response was, "No, it's closer to 100 million, and it's on a tear."

And just so you understand what those unit volume comparisons are, there were 120 or so million PC's shipped last year. There are about 100 million of those phones running Java. Those are growing a few 100% a year. So I assure you, within the next 12 months, we will out-ship PC's.

I think it's a testament to the developer community that's out there that's actually creating the content that's making that an engaging and interesting platform, and a testament to the team that actually had the vision to believe that a little, maybe ten-digit character display, three years ago, could evolve into something far more interesting and far more valuable.

On the card side, just so you understand what those cards actually mean, if you pop the battery off that phone and turn it over you'll see a SIM card, and a SIM card is basically a microprocessor smart card used for strong authentication in a phone, so there's at least 100 million of these in those phones right now.

But moreover, American Express' Blue Cards, the Military, every member of the American Military right now is carrying something called a Common Access Card, basically a SIM card, and they use it for building access and system access, making purchases at the PX.

These little microprocessor smart cards are becoming the kind of smallest atomic unit of network identity, and so those phones are beginning to be used for network authentication.

Telephonic, I was just telling you recently, they use those phones to log in the Web system. You go to a Web site; you ask to log in. It asks you for your phone number. You enter your phone number. Your phone displays a message saying, "You're seeking to log into X and such site; please enter your pin and click okay." You enter your pin. You click okay. You're logged into the Web.

So those devices, I think, are now becoming the hub of network identity, specifically related to the microprocessor smart card that's in the back of it.

We think that trend is going to continue. We see the growth in the Java card platform also to be growing or to be continuing.

I think there's been somewhat of a tipping effect in Java. At this point we're shipping, dramatically out-shipping the number-two competitor, which is probably Qualcomm and BREW. You probably saw recently, Qualcomm kind of ate their hat and actually licensed Java.

And moreover, and I think more importantly, this is the much-wanted Juan Dewar we were mentioning earlier. We have dramatically out-shipped to Microsoft Smartphone. I don't know what the unit volumes are in Smartphone. I don't know if you know.

A. Brennar A few hundred thousand.

J. Schwartz A few hundred thousand at best. So it is an abysmal failure, and for a couple of reasons: one, because they were late; two, because you don't want to wait a minute for your phone to boot, three, phones aren't supposed to crash; and four, they're not supposed to compete with the carriers that are deploying them. So for those reasons, we think the success and momentum of Java on the handset is just going to continue.

But all of that said, why is it interesting to Sun? It's interesting because if you look at the functionality delivered on those phones, all of it intersects with the Web Services infrastructure that we have continued to promote and drive on the server side.

You saw, recently, we were awarded a WS-I Board seat, which I think puts to rest finally, now, what is Sun's position in the Web Services world and I think really begins to bring, more effectively, the Java community into its involvement around the evolution of the standards in the Web Services world.

So we've had really significant updraft in the downloads of our Web Services Developer Platform. I think, at this point, our leadership, at least around the thought leadership of where Web Services is going, is undisputed.

Lastly, and I deliberately skipped over it, Java on the desktop. We have, as you might imagine, with Microsoft's announcement they were moving Java from the Windows distribution, seen a dramatic increase in the number of both consumers who were coming to get a Java run-time environment for their desktop, as well as an almost magnetic attraction by the PC OEM's to Sun, because they have now realized that if you're about to ship a PC and it doesn't have a Java run time and you're going to ship it to the heartland of America, that little Molly on her way to college, if she can't go play her favorite games at pogo.com or go chat with her friends at Yahoo or go use the AOL Instant Massager client written to Java, then she's going to make a phone call, and that phone call is going to sink the gross margin on the device that's actually shipped.

So the PC OEM's are, all of them, at this point, looking at licensing Java - we hope to have progress soon - mainly so that they can assure their customers the continuity of the experience they've begun to expect on the Web. So Java is used on seven million to eight million Web sites out there, everything from gaming sites to personal financial services, to weather, to sports applications. We think Java on the desktop is just going to get more interesting, especially now that we've got a compatible run-time environment on that desktop; we've got a very interesting opportunity to start interacting with that device that we just passed around.

What's next? All I can tell you is 100 million devices in the world is an interesting number. Those are not simply for consumers, although I would imagine right now the killer application for those phones is more likely Donkey Kong than Siebel. Over time we think that will change, not to suggest that consumers are going to start using Siebel, but people just want to have a reduction in the number of devices they have. And I'll give you a simple example, which I think is probably very tired for my staff.

But I was out - I had a very profound customer meeting. This happens every month or so. I was sitting with a company that operates mines; they pull ore out of the ground. I was having lunch with their CIO.

One of the issues these guys ran into is they have thousands of people who are out in the world who are all looking at things like, literally, settling ponds, where they dump a bunch of stuff in; the bad stuff to the bottom; the good stuff to the top; they skim it off and they sell it, but they have to monitor seepage because the government requires that they actually monitor the environmental impact of this stuff. So you've got thousands of people out in the world who walk around monitoring seepage.

And the way they do that today is they have a PC and they take it out to the settling pond, and more often than not they drop it in the settling pond, or it gets run over by a 300-ton truck or it gets filled with dirt and dust, and so they polled their workers and said, "How many of you have phones?" And you know what the response was? About 100%.

They thought, "The seepage monitoring application isn't particularly sophisticated, and there's no reason why," -- I'll show you another device -- "I can't actually display it on this. The display is large enough. There's a programming platform that allows me to use the same skill sets that I've got on the server side. All my monitoring applications are written to Java on the server side anyway. Now I've got a great opportunity to get device reductions. I can take the PC out of my infrastructure," because for a lot of these field force automation applications this has now become a pragmatic platform for the deployment of interesting Web Services.

So you're going to hear a lot more about that from us at JavaOne this year. I want to encourage all of you to be there. It's June 10th through the 13th in San Francisco, my favorite city. For more information, as well as registration information, you can just go to java.sun.com/javaone. So that's the developer.

What's going on with the CIO? They've got a bunch of different priorities. Probably the single highest concern, the single greatest set of issues they've got to work out really surround complexity and expense; they are related; they're the same thing. When I talk to CIO's, in addition to hearing about some of their platform concerns and how they're trying to consolidate, there's a couple of interesting trends.

One is that they are really moving towards shared services platforms. They are looking for fewer operating systems, not more. They want them to be cross-platform. They want to have common infrastructure. They want to have common skills.

If you think about the way middleware is licensed, this is one of the single biggest financial exposures for a company. You buy directories by the entry, mail systems by the mailbox, clusters by the node, app servers by the CPU, portals by the CPU, file systems by the terabyte; you're trying to manage all of this stuff; you've got a ton of lawyers who are involved; your terrified about which systems you're going to expose to which of your constituencies because you can't really monitor what's actually going where, and that is getting worse rather than better.

All that said, I think we recognize the x86 architecture and 32-bit platforms are growing in relevance to the enterprise. The good news is UNIX is playing a more strong role there as a result of what's happened with Linux, and also with Solaris and Intel, than it has before. And moreover, the opportunity that Sun continues to see, which is divergent, I think, and we recognize it from the way the rest of the world thinks, is that with IBM and HP largely abandoning UNIX and especially abandoning UNIX on 32-bit systems - neither HP, UIX or AIX run on 32-bit systems - we've actually got a consolidation opportunity. I want to reiterate we're well aware of the fact that that is not necessarily a popular view.

I think back in 1993 and 1994 when the world believed Sun had to adopt NT or it would be dead, I think we had an also divergent strategy, which I think we recognized that when that divergence exits, there's also an opportunity to make some money.

The reality for the CIO - and I'll come back to the platform discussion in a second - is this chart, which is a symbolic representation of what they have to deal with in their own environments. These elements over here on the left are pretty much the common elements of a shared-services platform in any enterprise.

If you are an enterprise, the odds are very high, 90% or above, that you're going to be running an app server, Web server, Directory, Identity, Portal Integration, Messaging, all of this infrastructure, all of it. So when you talk about an operating system, you can't talk about it absent all of those components.

Now when you look at that, these little dots represent the ship dates on this stuff, and these would be just the ship dates from Sun. You can imagine all the third-party components you're trying to get coordinated and all your own internal development you're trying to get coordinated, and again, remembering that, at least with respect to the infrastructure that exists on the left, you license each of these by a different meter: CPU, person entry, calendar, mailbox, cluster node, file system size.

So it struck us that we're actually pretty good at delivering on this bottom line here, which is an operating system. We do so once a quarter, haven't missed a beat in God knows how long, five years, six years, and it's a metronome-like consistency, a cadence of releases that provide updates that are backward-compatible and forward-compatible, that have the latest and greatest security modifications and enhancements and patches, and allows an enterprise to begin to plan how they roll their infrastructure out. It is remarkably simple. It is remarkably effective. It's what our customers really love.

The thought occurred to us, "Well, if they loved that, why are we bothering with all of these other release processes?"

What we announced at our Worldwide Analysts Conference was - this little guy on the left is Project Orion and he's putting an arrow through the heart of complexity - we are going to drive all of our middleware elements, everything you see on the left, the entire Sun ONE platform, into the same quarterly release process for Solaris, which means that the definition of an operating system coming from Sun on 32-bit systems and 64-bit systems is now all of that.

And so it is no longer, "Which kernel are you running?" It is now, "What is your Web Services infrastructure? What is your directory and identity infrastructure? What is your communications and collaboration infrastructure? What is your UNIX operating system? What is the skill set and technologies needed to keep it highly available?" This is a dramatic shift in the way we deliver software.

From an operational perspective, we'll take the bottom out of the complexity and therefore take the bottom out of the expense of actually operating these components. This is probably the single biggest shift in our software strategy in the last decade.

Now just delivering it differently, we believe that, because this will run on 32-bit x86 systems, because the same stack with run on 64-bit SPARC system, by integrating all these components together we can not only take a big bite out of complexity; we can also begin to deliver to our customers a methodology that will allow them to reap the benefits of a consistent delivery cycle around their own applications.

There's also a platform that we will make available and a process and a methodology that we will make available to our ISV's. So an ISV that wants to count on having an app server there or a portal server or a messaging server, or increasingly identity and directory - really the most pervasive Web Services out there - will now have a reliable infrastructure they can count on wherever Solaris exists. So we believe, again, this will be a radical reduction in operational complexity. What's the business impact now?

Operationally, what I just described to you in Project Orion is the single biggest shift in our strategy. What you'll see here will be the industry's single biggest shift in pricing because, again, you can, if you really enjoy the experience, go purchase this by the CPU, that by the entry, this by the node, that by the terabyte, etc., and we're going to not depart from our traditional mode of pricing. We will move toward a predictable mode of pricing. And what does predictable mean?

Predictable means you will know exactly how much the systems' infrastructure costs you. You will know what the meter is against which that measure will be taken, and you will be able to plan over very long periods of time and be able to deploy software in as great as quantity as you want without always having to come back to us and true up and figure out a way to pay us a bunch more money.

We have looked exhaustively at the Microsoft Licensing Agreement, the Oracle Licensing Agreement, the IBM Licensing Agreement, other licensing agreements; we have done the customer interviews with thousands of folks at this point to really understand, "What are your priorities? How could we really make this a lot simpler for you?"

While all the noise in the marketplace, I think, is beginning to get folks to concentrate on things that aren't necessarily adding a lot of value, we believe taking all the complexity and the expense out on 32-bit x86 systems, as well as 64-bit SPARC systems, we've got a real opportunity to totally drive a new economic model for the software industry.

So the odds are good that the pricing that we will be delivering will be metered against an auditable metric. We're not completely certain, but I will assure you it will be something related to the number of employees you have, and in so doing, we will enable large-scale enterprises to look at numbers where they don't have to take a guess at how many customers they have, how many CPU's they're using; they'll just have a very simple metric that we can all agree upon that will keep their costs both down, as well as more predictable, going forward.

To me, I think this is a massive change in the software industry and hopefully we'll put some pressure on our principal competitor to maybe deliver a more palatable licensing model.

Just to reiterate, what's our OS strategy? We are absolutely doubling down on Solaris. At a time when the rest of the world is abandoning UNIX, we believe that what customers really want is a uniform application model, with uniform infrastructure, that runs on Intel systems as well as 64-bit systems, but that's obviously not enough.

So in addition we will also deliver what is called Project Orion on Linux. We will do so through standard distributions. You will be able to run Linux applications entirely unmodified on Solaris, and we will make the commitment to our customers. We believe we can set the standard, not only for interoperability between Linux and Solaris, but even interpretability within the stack, which means, "If you don't like our portal server, take it out and put in Epicentric. If you don't like our directory, take it out and put in Novell. If you don't like our clustering, take it out and put in Veritas." So that's what's going on on the CIO side.

Lastly, I think we've probably filled your heads with enough perspective on N1, just to say we are in large-scale pilots now. We have close to about 20 large-scale enterprises running tens of millions of dollars of infrastructure over N1. We'll come back at another point and give you an update on this.

We've obviously announced our blades; announced the shipment of our provisioning software on those blade systems. To us, blades are just one form factor out of many that N1 will have to support. And we continue to add more and more systems - heterogeneous systems, HP systems, Dell systems, Linux systems, EMC storage, Brocade switches, Cisco routers - into the catalog of components that can be virtualized away and managed with an N1 infrastructure, but again, I don't want to spend a whole lot of time on that today.

I think, just to review quickly and then we'll open it up to questions, we are really pleased with the success. We have the numbers, we believe, to demonstrate that success around Java, both on the card space and the SIM cards, as well as on mobile platforms, as well as on desktop, as well as, ultimately, back on to servers.

With respect to the servers, we are continuing to double-down and focus on the things we know. Despite some of the movement in the marketplace, it's looked at as a curious investment. We are absolutely convinced that the delivery of a common platform across 64-bit and 32-bit systems that contains the entirety of the Web Services infrastructure necessary to run a 21st-Century business is the right strategy, and that Orion really has an opportunity to go drive complexity and expense down at a time when CIO's are intensely focused on that.

Then finally, N1 is the right model to begin thinking about visualizing all of this away. I think, potentially, one of the interesting areas for us is to begin looking at to what degree these two things ultimately intersect, which is "If you have written applications, why haven't you written them to be more inherently provisionable?"

With that as a teaser, those are all the slides I've got. If I could ask my folks to come join me, we can open it up for Q&A.

M. Petry Thanks, Jonathan. At this point we will open it up for Q&A. Hadley has the microphone. For the folks on the phone, if I could ask the operator to start queuing up questions. For the folks in the room here, if you could wait until you get the microphone to ask the question so the folks on the phone could hear the question, that would be great. Also, when you do ask a question, if you could state your name and affiliation that would be much appreciated.

CNET John, you were talking about the wonders of volume for attracting developers; great with Solaris, SPARC servers; great with Java phones, but are you convinced that the volume of x86 Solaris from Sun is going to exceed the volume of x86 Linux from the collective server market?

J. Schwartz I think when you look at the Linux marketplace, you really have to look at really two companies, which is Red Hat and SuSE. So Linux, as something you can't see or touch, is kind of difficult to quantify.

We've got a couple million copies of Solaris on Intel now, so a couple million that are out there. Will there be more copies of Solaris on Intel than there will be of Red Hat or of SuSE? It's going to depend.

I think for generalized infrastructure my bet is, given the way our volumes are going, we will be headed in the right direction. I'm not sure we look at it as a race; I think we're more interested in the impact on Windows.

I think to the extent that Linux has made its way into kind of transparent systems, systems that aren't actually exposed to programmers such as network security appliances and intrusion detection, those unit volumes will likely to continue growing, and to the extent that Linux ends up, or Red Hat specifically or SuSE specifically, on set-top boxes or consumer devices - this phone that I'm carrying right here is actually a Linux phone. It's running Java, and Java is the entirety of the user interface and programming model. But I think those unit volumes are going to be really dependent upon which of the elements you're actually talking about. So generalized server-side infrastructure, it will be an interesting comparison, I think, a year or two from now.

For devices that are out there, don't expect to see Solaris running on one of these any time soon, unless you want a scaleable, multi-threaded, large-scale Web Services platform on your phone.

John, do you have any comments that you want to add?

J. Loiacono I think we're very confident in the fact that where we're going with Solaris right now on Intel or, actually, x86 platform, I should say, we're very confident and we're glad that you're going to see some announcements here over the next several months talking about both wins as well as partnerships that we'll be announcing to show momentum in that space.

J. Schwartz If I can leave you with one thought there, I think we recognize and understand that we are not the only supplier of x86 systems, and that it'll probably behoove us to ensure we have a good relationship with other x86 suppliers who would be in a position to deliver Solaris. We believe that feeling is mutual.

RFG My name is Michael Dortch. I'm with The Robert Francis Group (RFG). I'd like to ask about more details, if available, and the time line for when enterprise customers might have the option to talk to Sun about some of the licensing alternatives today.

J. Schwartz You will start seeing large-scale licensing deals that Sun is doing, folding out literally this week - we were expecting it to be before this - next week and the week following, and just on and on and on.

I've probably done a couple-hundred customer visits in the last 90 days. Orion has been a factor in every single one of those visits and it's been very well received, but let me give you an example of where it was not well received and what some of the challenges we're going to face might be.

I was with a big investment bank recently, and in a room I had the folks who do Web Services infrastructure, largely app servers, portal integration technology, directory and identity; folks who do security and access and provisioning systems; folks who do communications and collaborations, e-mail systems, instant messaging, calendaring, on and on; and then the folks who do UNIX, and they have Solaris and Linux and HP UX and AIX. Then I gave my big Orion pitch and all of them universally said, "We hate that. That's a stupid idea. That's just brain-dead. Why would you ever do that? We're going to take all the components we want."

Then the next day I went and met with their CIO. You know what he said? "I love it." So I think this is a different way of thinking about operating infrastructure.

I see CIO's looking to consolidate their purchases. There is an enormous consolidation around the brands that matter in the enterprise, and the platforms that matter in the enterprise. We believe we have an enormous toehold with Solaris. I think we're late to the market - we'll freely admit it - in delivery x86 systems, but now that we're there and in force we believe that that consolidation is something that will benefit Sun.

RFG A couple of quick updates, if you can provide them. Can you say anything useful or interesting about United Linux and/or the Liberty Alliance or are they both going the way of the late lamented Project Monterey?

J. Schwartz Liberty is out of the press and into the customer base. I was just with a large portal company that's looking at doing yet another big deal with a big telco, and guess what? They've got to get their identity systems to talk to one another; what else are they going to use? Passport? I don't think so.

So I think Liberty is just continuing to creep along, principally in looking at single sign-on behind firewalls, because that's where a lot of the pain has arisen. The most fundamental Web Service that's out there, despite all the hype around Web Services, is logging in. How many different places in your enterprise are you actually logging in and trying to actually get access to a service or a completed transaction?

So providing uniformity and shared single sign-on, federated across … is challenge number one. That's what we deliver today. That's where we have customer wins today.

I think, going forward, at the point where businesses are able to get out of the mode of thinking, I think, historically, that each has their own customer base and they look at syndicating relationships across the Web, Liberty will play a very strong role in part because that is the standard that exists; the other one just doesn't. So, John, do you want to comment?

J. Loiacono On the United front, I would just say, as Jonathan alluded to, we are going to be supporting standard distributions moving forward. I can't name the specific ones. I'll just tell you there are probably no fewer than two and no more than three or four.

So where United plays into that, of course, is the fact we like the concept of unification. We're into, again, multiple; we're not going to pick a single vendor by any means. So moving forward, we want to see multiple options there.

So where United plays into that is, of course, more of the marketing and go-to-market side versus the standard distribution. SuSE is a driving force in that space, for the most part. So we'll have a relationship there as well.

J. Schwartz I think we also recognize that enterprises are increasingly recognizing that they are writing to a distribution. They are no longer writing to Linux because what you write to Red Hat Event Server will not run on SuSE Enterprise Server. And so in some sense, competing with that which doesn't exist and that which is free is really, really hard, and now that some of the clarity is coming into the marketplace around both licensing terms as well as the operating realties of running Red Hat or SuSE, I think we recognize we've got to adapt to that as well as potentially deliver that to the marketplace.

M. Petry Great. Thank you. I'll take a question from the phone, please.

Computable I work with the IT weekly Computable in the Netherlands. I have a couple of questions on Orion. For the other UNIX distributions, how will Orion deal with them?

J. Schwartz You mean Hewlett Packard and AIX?

Computable Yes.

J. Schwartz I think we'll look at them as market opportunities. I will tell you that what I've seen is if your UNIX doesn't run on 32-bit systems, we're beginning to see a bit of a freeze on customers' willingness to invest in them. So I think we have a competitive opportunity, obviously, because those stacks don't exist on those other UNIX's, but I think we also recognize that our customers don't all run on Solaris, and therefore we've got to do our best to support them.

I think we are looking at this less in terms of what's the competitive advantage we have against HP or IBM and more in terms of what can we do to keep costs and expense down for enterprises, and we obviously want to start with where enterprises are and not necessarily drive them to where we want them to be.

Computable Thank you. The synchronization that Orion is dealing with, how often will that take place?

J. Schwartz Once a quarter, just like the Solaris release today, and just so that we're clear, what Solaris offers is backward compatibility and forward compatibility. So what you have written to Solaris 7.0 and on, I think, will run on what we deliver, and that is a couple years, too many years, of platform consistently. On the other hand, that's, in part, why customers love Solaris.

Computable A couple more questions here.

J. Schwartz Why don't we do just one more question and then I think we'd like to share the floor.

Computable Can you name some other vendors or what kind of other vendors are falling into this fold of Orion?

J. Schwartz Other ISV's?

Computable Yes, that are working along with you now to do the same kind of scheduled releases.

J. Schwartz We're talking to pretty much everyone about it. We've been very public about it. So I'm not sure we'll have any named partnerships with respect to what we distribute to customers. I think you should expect to see vendors begin aligning their releases around our releases, but I'm not sure there are going to be any formal announcements around that.

Computable Thank you.

M. Petry Michael, you have a question?

Jupiter Media Michael Singer with Jupiter Media. Jonathan, once when we talked about Orion you said that of the three licensing models that metering was your least favorite, and I was wondering if you had a change of heart on that at this time.

The second thing is being able to swap out certain components of the system for your own preferences, will you still be charged full force for that particular licensing or is it going to be also not only how many employees you have, but how much of the Solaris base that you're using?

J. Schwartz With respect to the former question - let me leave the former question alone for a second.

Metering, just in general, I don't buy it, mainly because I don't know a CIO around that wants to tell me how many customers he has, how many CPU's he has, wants to agree to a mechanism to count some arbitrary unit of consumption. You know, what is an entry in the directory? Does it depend on the number of schemas you have? Does it depend on the number of items and attributes in the directory?

It's a bizarre way of trying to get, ultimately, what customers really want, which is more predictability and a capacity to ration their consumption. So we'd rather take the cap off of rationing and say, "Use as much as you want;" bear in mind, the marginal cost of delivering another entry in the directory is pretty low for Sun, and really let them run their business as they see fit.

I think predictable is what customers want. I haven't found a lot of appetite for metering other than in the press, to be blunt, because it's such a sensible concept because that's how you make phone calls today; you meter based on how much you consume. No CIO really wants to get a monthly bill that says, "Oh my God, my people all made huge entries in their directories this month. I had no idea I was going to have that." They're just as open to surprises.

So will the price be the same if you take out a component of Orion? Yes.

Jupiter Media The follow-up is with regards to working with other partners, how is the process coming along with JBoss?

M. Bauhaus There are two things that are important here. One is that for the Java industry in general, and certainly it applies for J2EE as well, compatibility matters a lot. There is no Java industry without compatibility. So we're focused on making sure that we have that.

We've offered JBoss the opportunity to make sure that their implementation is fully compatible and obtain a license that basically certifies that. The ball is in their court.

Just in general, in the industry, it's compatibility that makes all the app server vendors and, frankly, Java Web Services a reality, and we want everybody to be able to work at that level.

J. Schwartz I think JBoss needs to figure out are they a for-profit business or a not-for-profit business. I think it was very easy for Sun to deal with the Apache foundation. So with respect to JSE's and servlets we gave them the compatibility test and therefore they are compliant and we had a very positive and productive relationship, and we have extended that offer to JBoss. I think we're just still trying to figure out are they in business to make money or are they in business to promote a free good.

M. Petry Thank you. Can I take another question from the phone, please?

Open Sys. Advisor Nina Lytton from Open Systems Advisor. I was wondering, Jonathan, when you say that the other vendors have abandoned units, it became clear to me today that you're specifically referring to the 32-bit platform.

J. Schwartz Yes.

Open Sys. Advisor So my question about your opportunity is, when you're even out there talking with the 100-odd customers in the last little while, where do you sense that the financial opportunities for Sun and the value opportunities for customers are with Solaris on the 32-bit platforms?

J. Schwartz I think when you look at our pricing on Intel, the price of the kernel is only one price. The added-value elements that have really become to define the infrastructure on which you run a business are Web Services, directory and identity, messaging, collaboration, administration, resources management, all of that. Going into a comparable stack, you can just do the side-by-side comparison yourself between Windows data center and us and some collection of other stuff that you try to build up on another x86 system. So I think when you do that analysis you'll see a very rich vein of margin opportunity for Sun.

Secondarily, maybe to respond specifically to your question around "Do we believe we're going to make a ton of money off x86 hardware," the margins are still interesting; they're not as interesting as the software; the software is a competitive opportunity that we have.

If you go to any other vendor of x86 systems over the Web - and I encourage you all to go do so - they'll configure it with your favorite app server, portal server, directory, calendar, messaging, all of that infrastructure. Do the math; look at how much it costs. Just bear in mind we don't have to pay for any of the software and we will go to the same Asian suppliers to get the motherboards that go into the x86 systems.

We believe that's a healthy business opportunity. We believe it represents a lot of margin opportunity. We're shifting a lot of the manufacturing to others. And I think we've just got to stay focused on providing customers what they want, which is a uniform platform that spans both microprocessors.

Open Sys. Advisor What applications specifically do you see them buying those systems for, from either Sun or from another x86 box provider?

J. Schwartz What applications are customers using x86 systems for?

Open Sys. Advisor Where you see that you have the opportunity, what particular areas of the infrastructure.

J. Schwartz Generalized enterprise infrastructure, Web Services infrastructure, specifically. I think there's a lot of it and this is a much longer discussion.

If you think about, remember all the original Web servers companies that were out there - there were probably 50 of them, and then they consolidated away, and now there's just basically three left: Microsoft, Sun with the Sun ONE, formally NetScape Web Server, and Hitachi. Then people started writing applications on their Web servers, so all of a sudden you needed transactional logic and containers and a lot of other infrastructure and therefore an app server existed, but it kind of subsumed the Web server infrastructure. And then people starting worrying about the presentation layer of their application, so they created this thing called a portal server, which contains the Web server infrastructure.

So when you put in all in one place, what do you end up with? A lot of redundant infrastructure. You can put it all in the same system. You can begin looking, at a systems' level, of how you optimize that infrastructure.

One of the, I think, enormous competitive advantages that Sun has is just take a look at the folks that are building the systems at Sun. We can all sit down, as we did yesterday, in a very boisterous, as Alan referred to as a "fracas", to really look at what are the ways we need to modify our systems so that they are more efficient, more coherent, more readily operated, more effective in deploying, very quickly, services that are being generated in a sea of new devices; that's exactly the opportunity that Sun has. We're probably the only company other than Microsoft that has that breadth of technology to be able to make those decisions.

M. Bauhaus I could add to that. Nina, it's Mark Bauhaus speaking. The people are writing now to the platform sitting above the operating system. So you write to Java Web Services; you write to this combination of Web server, app server, portal, identity, directory and integration services that are all part of a complete thought, if you will.

The deployment decisions are where you decide to use x86, whether it's 32-bit, 64-bit, which operating systems makes sense and, increasingly, that choice is more transparent, and N1 makes it even more so.

Open Sys. Advisor With Orion, you're just helping people take cost out of what they're doing anyway.

M. Bauhaus That's exactly right. So for a developer, they don't really have to think about the operating system very much at all. They have to think about it as they think about how to deploy and what's the most effective environment from a scalability, manageability perspective, given the kind of function that they want to do in that tier or that part of the application architecture.

M. Petry Thanks, Nina. I think we have a question from Richard in the front.

J. Schwartz By the way, if we get one more question on enterprise infrastructure, then we're going to talk about consumer and mobile devices because we dragged Alan all the way up here and he's like a fountain of knowledge on this stuff. Go ahead.

Financial Times I've got the infrastructure side.

J. Schwartz How many of you have phones? All of you. Good. How many of you use them at work and for work-related things? We'll come back to that.

Financial Times Richard Waters, Financial Times. Two related questions to what you've been talking about. One is the economics of all of this. Some of the systems' companies had a gross margin of whatever it is, 45%. When you look at Solaris on x86 and you look at Orion on Linux and so on, what is that going to do to your blended gross margin over time? Presumably it's going to come down?

J. Schwartz I disagree that our gross margin on the bottom end of our business goes down. In some ways it's easier to introduce new software infrastructure on low-end systems than it is on a 15k that's running somebody's reservation system.

So in some sense our margin on our low-end systems - I mean if you look at the combined margin of a Microsoft/Dell product out in the marketplace, there's interesting margins there; it's just that you have to look across both Microsoft's business and Dell's business, and the same thing applies to us because we'll supply both. So I do think we have margin opportunities there.

We've continued to believe - and I think the marketplace is now beginning to bear out - there is no such thing as a free lunch. With Red Hat's new pricing level they basically said, "There's no more free. There's no such thing as a free thing from Red Hat anymore." So Linux and free have gone away. And the problem with free is that gross margins are terrible, and in the new world order, if you look at the price of Windows, the Data Center Edition or Advanced Server from Red Hat, there's a lot of money in there because these things are complex. They actually take time and money to go support.

So I don't know if you have the most up-to-date pricing from Red Hat but they're a for-profit business; they're trying to make money. They're doing the right thing, which is trying to establish a price for the value that they deliver, and as that happens, there is margin opportunity there.

J.Loiacono So if I could answer slightly differently, and that is that - this goes back to Jonathan's point he made in the last question, which is people keep saying, "How can you compete with the low-cost providers and the perceived low-cost providers?"

The answer is that if you just want to get a box and an operating system you're probably not going to come to Sun, and we're okay with that; that's not our business. We're a systems provider. We'll take the operating system on a piece of hardware and then we're going to give you a stack of software, currently called Orion. That's going to all be integrated, sold, serviced, and supported on a worldwide basis. That's our value proposition.

If you buy a box with an operating system - name your favorite one - eventually you want to do something with it; you want to run an app server, a portal server, Web server, directory server, etc.; you want to go do something.

So our value is that if you go buy one of these low-cost producers, or perceived low-cost producers, then you've got to go by an app server from one company, a Web server from another company, a directory from another company, and then you say, "Is this low-cost producer going to integrate that for me and then ensure that it works together? Are they going to service that in a worldwide basis?" That's where the question… and that's where Sun's value comes in and that's where we can monetize it.

J. Schwartz I think one of the questions I have - and then maybe we can start talking about consumer devices - is there are definitely applications, which don't really require an OS anymore. A perfect example of that is a render farm. You're going to run a very trivial application and you're going to run it across 500 servers; what OS infrastructure do you need? Not a lot, a tiny networking stack, a little IO, maybe a file system. In that marketplace I doubt anyone is going to make money; I don't see how you do it.

Ultimately the dominant provider of that system is going to be Intel, more so than any of the OEM's that are out there, because they will have no value to add other than a pretty bezel and a new logo and a different skin on the top. And even Intel, I think, is going to have a hard time doing that because people who deploy render farms tend to buy the cheapest things that's available, not the latest and greatest, just the cheapest, which means they'll go to eBay; they'll go to wherever; they'll go to Taiwan and get the latest and greatest cheap thing that you can get for $50.

I think you really have to look at the spectrum of applications to which infrastructure will be put. Generalized enterprise infrastructure, which is our business, is very different than intrusion detection or just a render farm, so I think you really have to look at what's going on.

I think the margin opportunity for us on low-end systems, because we have an established software franchise, is really interesting, and I think if you are another PC vendor, another x86 vendor and you don't have a software stack, you're going to be paying someone else for it. That is a margin opportunity for Sun.

Financial Times I am going to ask one Java question. You mentioned Java on the desktop before and the fact that you think Sun desktop manufacturers are interested. Do you think you might have a deal before the Appeals Court ruling or do you think all those people are going to wait to see what happens with Microsoft?

J. Schwartz That's a good question. I'm not sure. There's no linkage; that's the main thing. There's no linkage between them. There's interest independent of what happens in the courts. It doesn't matter.

Let me give you an example. There's a media company that happens to be a PC supplier, and you'll figure out who it is. I think the problem with being a company that depends upon a standardized infrastructure for the delivery of gains or media, which is what this one particular company does, realizing it's just been pulled out of your distribution, which means your content is now at risk, your business is now at risk, almost independent of the Appellate Court decision, they're probably going to find an alternative supplier of a Java run-time environment.

But I think your question is a right one, which is, yes, there is a lot of, "Well, we'll just kind of wait and see what happens next." I think inevitably all those companies are going to decide, "I've kind of had enough. I don't want to deal with the issues. I want to have an integrated, common, safe infrastructure that's compatible and I can just run my business."

Register Ashlee Vance with the Register. Sorry to go back to the enterprise stuff, but in Orion, I'm just not clear; you mentioned Sun, Linux, SuSE and Red Hat. Can you tell me how they're going to be distributed on all three? Are you sticking with Sun Linux?

J. Loiacono With Sun Linux, we will not be supporting the customized version of Sun Linux moving forward. We will move to entirely standard distributions. So I haven't selected which ones they'll be. I told you there would be two or three, four at most, but they will not be Sun Linux as you know it, meaning they custom take the standard distribution and modifying it.

RFG As we all know, there are no interesting consumer devices that are network-enabled without there being some back end. But I don't want to go there yet. I do want to go there.

The problem that we see, as analysts of Robert Francis Group, is that the consumer device manufacturers keep coming out with these really cool devices, but the carriers don't really understand what the hell to do with them. And when you talk to carriers about data services, they tend to show you things like being able to wirelessly watch a move while you're riding around in a limousine, which is not really a value-added service, as far as we can tell. So can you talk a little bit about how Sun can help make the carriers smarter to help everybody sell more consumer devices, which can only be good for the enterprise-oriented back end?

J.Schwartz That's a good question. We're doing a number of things. Sun is not a consumer company, but there are small groups within Sun that spend a lot of time actually talking to consumers and big OEM's and big carriers that deal with consumers. What we're trying to do is bring some of the technology, some of our expertise in solving complex computing problems, and go to the carriers directly and help them enable these services that leverage the power of the network and the knowledge they have of the customers and make it a lot more attractive to the consumer first, and then we're going to take that to the enterprise, also. So we are doing it. We're involved with companies like Vodafone, like Sprint, companies in Japan. We're working on it, but it does take time because it is a large, complex ecosystem.

But you've seen things on Vodafone live a couple of days. They talked about one million consumers. We're in there working with them.

A. Brenner There are several fronts you have to work at once to solve the problem that you've raised. One of it is making it easier for the carriers to acquire content that they can deploy, which they know will be readily modifiable. Another one is to actually work with the content providers to help create interesting content that will be attractive to subscribers, will be well-suited for the variety of devices that are out there and the limitations that those devices have, the kind of devices Jonathan passed around earlier. That's a very clever design, I'm sure you all know this, but you don't see something like that very often.

We're ten years into cell-phone design, so it just takes a while for those things to pop up and cross people's imaginations and get effectively executed against.

The business process has to be streamlined; it has to get a lot easier across the board for carriers to spin-up a new service and deploy it, and that's not just the phone issue; it's a carrier infrastructure issue all the way back to the IT, all the way through the intelligent network; these things are all happening in parallel.

The good news for Sun is, because we are the provider and leader of Java technologies and in the network, we get tapped on the shoulder regularly to help create the systems to do this. Maybe one of our most striking examples was the rollout of Sprint PCS Vision, which was a system that Sun worked day in and day out. We had people living at Sprint, designing the infrastructure for PCS Vision. So we spend a lot of time these days helping carriers to sort this out.

J. Schwartz For those of you who are calling in from Europe, you have, I think, a more positive perspective on carriers because they have executed and delivered on a far more sophisticated infrastructure than we poor heathens in America have to live with.

The infrastructure that exists here - I carry these very interesting and engaging devices, but this is my phone. It has a, maybe, 15-character black-and-white display with five lines because in this country it's very difficult to get anything deployed to you that's any more engaging. So I think Vodafone, NTT DoCoMo, Sprint, AT&T Wireless, others have really begun to deliver far more engaging content, and maybe I'll leave you with one thought.

The content and ISV's that the carriers want today are in Hollywood and Redwood Shores, like Electronic Arts. They are dominantly looking at ways to now get onto these handsets with intersecting content. Am I going to watch a move there? No. Am I going to play John Madden football? Maybe, odds are good. Now is that going to help me as an enterprise? No, but is it really going to take a lot of work for salesforce.com to expose their sales force automation contact list on there? Trivial, we can do it now. So I think it's beginning to come together.

A. Brenner I'll add one more thought. So last year we had, at the end of 2001 we had about 25 million or 35 million phones out there. At the end of 2002 it's 75 million or 100 million, depending on how you count. So that's a huge lever.

The number of content developers that are going to be interested in a 35-million-unit market versus the number that are going to be interested in a 100-million-unit market or a 200-million, it doesn't go up linearly; it goes up geometrically, and that's what we're seeing right now.

We are seeing, pretty much, an avalanche of new little start-up companies or large content companies engaging in the market, starting to build applications that wrestle with the limitations of the user-interface vehicle, but take advantage of the fact that people carry them on their person, the fact that they have very interesting graphical displays. Things are starting to happen. I think you'll see a lot in the next year of new and very interesting services coming out in this country.

ZDNet Sorry to go back to the enterprise. A lot of your customers you're going to, I assume they already have an infrastructure in which they paid for Web servers, portal servers, and everything else; so when you go in with Orion and say, "It's an all or none buy," what kind of reaction do you get because if you're trying to push somebody out, unless it's an existing Sun customer, how do you go in and just say, "We can't afford to take the whole stack today"?

J. Schwartz I was just in with a carrier and we do tend to be a little telco focused, which we would talk about in a moment, and they were about to roll out a service to 100 million consumers, 100 million. They were going out and they were going to get an app server from a big company. They looked at the quote they got; it was $45 million. Our Orion quote all in, all up, everything, wasn't [sic].

ZDNet What was it?

J. Schwartz Much less.

ZDNet How much?

J. Schwartz A lot less. Enough to the point that it made it an interesting proposition.

ZDNet If all they wanted was the app server, they actually got everything else for free, theoretically?

J. Schwartz Yes. And moreover, I also want to be really clear - and Mark can talk about it as well - we're not abandoning our traditional mode of pricing. If you want to buy by the CPU and by the entry and by the identity, knock yourself out.

We believe that we are incenting the behavior that CIO's have asked us to incent. We want - and to be really clear here - a shared services platform, shared Web-services platform: shared Web Services, shared identity services, shared operating-system services, shared service that we can run at scale.

So when I go to a cable company recently and they're serving three million consumers with e-mail and calendaring and instant messaging, and they then turn around and look at their internal infrastructure where they have 50,000 employees and they're running Microsoft Exchange, they're doing it for three million people on one set of infrastructure and 40,000 people on another set of infrastructure; which do you think they're going to re-purpose for the other? You raise exactly the decision that a CIO has to make.

So we understand fully that the value equation here is really simple; it goes like this. The value is equal to the value of a shared-services platform - I can't estimate that for a CIO, but most CIO's are good at calculating - minus the cost of transitioning, minus our price. So I control one of those metrics.

M. Bauhaus I understand part of where this is coming from. We've gone through a period in the maturity of the Web where, basically, artificially created groups of features have been defined as marketplaces. Jonathan was talking earlier about the evolution of Web servers to app servers to portal, and now identity and integration become part of the mix as well.

Well those are really artificial, very early-stage markets that have been created with excessive economic rent. That's going away in a hurry. People want a complete platform. They want it to be fully integrated. They want it integratable so they can take out components if they want to and so on. There's plenty of room in the market for that integration at much, much lower economic rent, if you will.

Just a couple of days ago we announced the Sun ONE Web Services Platform Developer Edition, which is really what Orion developers can be using today to develop applications to be ready for that integrated stack. It's over $37,000 of software if you would have bought those components independently from Sun, much, much more if you would have bought it from our competition. It's a $999 street price for a developer. It includes Directory, Identity, Application, Web, Integration Services, our Studio, sample applications, all at one install, one price.

ZDNet How does that compare to, let's say, WebSphere or Visual Studio or BEA or anybody else who's out there trying to sell the either the welded-shut or the proprietary-open interface?

J. Schwartz I think that you really have to look at two other vendors. How do they compare against Microsoft? Go find out. You can do the math. How do they compare against WebShpere? What's their directory server? What's their identity technology? Not there. So if you want to start hand-assembling stuff - I guarantee you right now if you go out in the OpenSource world you can open LDAP, Send Mail, Apache, Tomcat, Linux, DBN - you can go put all this stuff together tonight, half an hour, maybe a little more, it depends on your bandwidth, and then you have to do work to make sure it all works together and it qualifies together and it inter-operates in a common installation and that's the next two weeks.

M. Bauhaus And WebSphere itself, even without the components that they're missing, you've got six to ten different products you have to buy at separate prices, install separately for well north of $37,000 of cost for a developer to get going.

J. Schwartz So we believe that - and this is a fundamental assumption for everyone who is a part of Sun, and certainly a part of the software team - we are a systems company; we create systems. We will never sell a bare metal box and expect to make any money at it, anyway; we're not good at that. Dell is great at that. Dell is fantastic at shipping the systems that they build, in part because there's not a lot of differentiation.

Where we are good is at looking at very complex multi-component systems; looks like work stations to begin with. We took somebody else's microprocessors or somebody else's motherboards or somebody else's OS, somebody else' sheet metal, and built a workstation and delivered it. We did the same thing on the server. We're doing the same thing in the software stack.

Why is Java becoming so successful? Because it's a programming platform, because it's a service platform, because it has an architecture that coherently can really take the complexity out of the system. To CIO's and to our customers and our OEM's that's a huge amount of value. That's what we really want to make sure that we're focused in on.

M. Petry Thanks, Dan. I want to invite the folks on the phone to ask questions, if you have any, and then we'll get to Michael in a second. I wanted to see if anyone on this side of the room had questions.

CNET I had a couple of Orion questions. When you're talking Orion pricing, is it one price for that entire stack or is it per module? I mean obviously it's by employee and the company or whatever kind of gauge you use, but does that vary according to how many of the modules you've used?

J. Schwartz No.

CNET When you're talking with this different Linux strategy is this going to be something where it's kind of an Orion appliance or is this going to be an open-ended Linux where you're working with a whole bunch of other ISV's to make sure that they're on board as well, for example Oracle unbreakable Linux on Sun?

J.Loiacono I think the initial plan is that what you'll see from us this year, just to give you some phasing out, we talked about the whole packaging. By the end of this year you'll see us take the major components. Jonathan showed you that slide that had 30 or so items on it. We're going to take the top five to seven to nine items off that list, the key ones, and that will be the first version of Orion. It will have a level of integration with it, and then that will be developed on all three platforms.

In the process, we're talking to third-party vendors - the Oracles, the Sybases, the other companies that we could actually do a simultaneous link-up. But what you'll see from us is our shipping our products initially. We haven't yet gotten into how we're going to actually announce the third-party applications and how we're going to integrate those over time.

CNET Specifically regarding Orion, are you going to be doing ISV partnerships with Linux?

J.Loiacono It's pretty dependent on the operating systems. So when we're talking about Orion, it's all the software suite above the operating systems, so they will run on both platforms, Solaris and Linux.

CNET Right. I understand that. Let me phrase it a different way. How is your Linux strategy going to differ from, say, Hewlett Packard? Are you going to be working with this huge collection of outside ISV's or is it going to be pretty much Orion?

J.Loiacono I'm going back to standard distributions. Therefore, whatever runs on the standard distributions runs on my platform, by default. I don't have to go attract Veritas to my version of Linux because it isn't my version. So whatever runs on a standard distribution therefore runs on Sun's version of Linux, if you will?

M. Petry Can I have the question from the phone, please?

CRN Can we have a Mad Hatter update? I hadn't heard much about that lately. I was just wondering if that's going to be out. I think it was supposed to be out in June or July. I was just wondering about that.

J.Schwartz We're still marching ahead. I think the thing that's interesting about the desktop Linux marketplace is, unlike the server where it's really kind of down to one or two companies, the desktop environment is much, much more diverse, and I think DBM probably has a more significant share than probably anybody else out there.

So that diversity, I think, is healthy. I think that diversity is yielding the Motorola decision recently to kind of kick … out and run Linux on their phones with Java as a run-time environment, both the programming platform and user environment, and we're fully intending, still, on shipping a desktop by mid-year.

JavaPro Dan Ruby from JavaPro. We talked a lot about complexity on the deployment side. It's a big issue also on the development side. I'm curious to hear what we should be looking for in terms of tools, particularly maybe in that space for visual programmers, higher-level business developers, and then maybe this also is a little bit of a consumer-side question because are there tools for J2ME that we also ought to be looking for?

M. Bauhaus Let me start and then you can jump in on the tools. First of all, this Web Services Platform Developer Edition that we announced last week was a big step forward because instead of having to spend days and days trying to get all these pieces together, you've got it all in one install; you are set up within a half an hour, including Web sample applications and guidance about how to use all the features of a full Web-services platform. This is new for the industry, I think, so it's a step forward.

We've oriented our future engineering towards that, increasing iteration and ease of use of these many features. There's quite an extensive range of things you can do now with Java Web-services environment.

J. Schwartz I guess a couple comments on the tools side of this. There's kind of an interesting bifurcation in the marketplace around the small- and medium-business developer, and the most interesting to me is salesforce.com, which is, for those of you that don't know it, they're this fantastic business, a really vibrant little ASP; there are some profitable ASP's out there, and they provide, basically, sales force automation for non-IT shops for small and medium businesses.

I think among their biggest customers is Kikkoman, the soy sauce people. Why? Because Kikkoman didn't want to try to do a big sales force automation rollout, what they did was go to salesforce.com and said, "How about you track our pipeline and our channel?"

So was Kikkoman doing a lot of software development when they made that decision? No. They went through a URL, www.salesforce.com.

Do we believe that we need to continually simplify and make more accessible the Java platform? Absolutely.

M. Bauhaus And we are, so stayed tuned for J2SE 1.5, where there are a number of enhancements and simplifications coming. Lots of companies working together to make that happen.

J. Schwartz I think the tools on the J2ME side, I think among the more popular are Photo Shop and, Alan, maybe you can talk.

A. Brenner The Sun ONE Studio Product is the mobile edition, which includes support for J2ME software development. The Sun ONE Studio Mobile Edition is built around some core technology called the Wireless Toolkit, which is also something that can be directly downloaded off the Sun Developer site. I think we've had about half a million downloads of that technology, and that's actually often used by developers just to get started in developing their applications.

That technology has also been integrated with, I think, on the order of 20 ID and STK products from the standard IDE companies, like Borland, but also included in STK's that come from the handset OEM's that are trying to attract developers. It's also been customized for use by carriers that are trying to build their own wireless developer programs.

There are also a number of other technologies that are in the pipeline that you'll be seeing news about - say, in JavaOne time frame - to simplify the development of the application end to end, so stay tuned.

J. Schwartz You shouldn't be surprised - and this is a bit of a twist, again, to force you back into talking about the consumer space. You shouldn't be surprised to see carriers as major sponsors of JavaOne this year. If you think about that, that's for a reason. The vibrancy of their business depends upon the health of the developer community, our developer community. You're not going to find a carrier supporting a Microsoft developer conference. So those folks are billions of consumers across the planet. That is the marketplace for the content, the services and infrastructure we provide.

So again, despite the telcos have a little bit of debt that they need to go work off, at the end of the day they're in a very, very primary position because they own your network identity. You have a relationship with them. You tend not to churn these more enhanced handsets. Why? "Because my life is in there. I buy my movie tickets. They have my buddy list. I've got my addresses. I've got my voice mail. I've got my follow-me services." So that's definitely the way we're viewing the market.

Jupiter Media I do have a mobility question, so we can tap into that, but firstly, a quick clarification on an enterprise question.

As a clarification for Solaris, in your documentation, in your slides here today you're talking about doubling down on Solaris for Intel and SPARC. Does this mean that you're identifying Intel as your preferred x86 partner? What about the other x86 partners?

As far as the mobility question, the SIM chip that Jonathan alluded to in the beginning is an identifier getting into portals and such. Are you seeing that as the future of what we're talking about, as far as your identity vision, or is there something else? Are you working with, say, Intel on their new phone shift, and where does that fit into with your Motorola strategy as well? What about PDA's? You didn't say anything about PDA's.

J.Loiacono Let me do the first one and get it out of the way because it's pretty easy. We don't have a preference on chip sets for Solaris, or for systems for that matter. So you'll see, I think, today, the LX50 is an Intel-based device. However, down the road, whether that's Intel or A&D or others, we have not made a selection on the preference, by any means.

A. Brenner Let me just answer the SIM chip mobility question first, and do it first by way of anecdote.

We have done an implementation of authenticated access for the Defense Manpower Data Center at the Department of Defense. This is an organization that is responsible for all of the civilian employees for the US military and today deploy a common access card based on the Java card technology; at Sun we have a badge like the common access card for the military; it's got a chip embedded in the plastic. The goal for that system, right, is to both increase security without making it more difficult for people to get at the services on the network.

So for example, based on the deployment of the common access card at the DMDC, they're working hard at eliminating their corporate firewalls. Now how can they do that? The way they can do that is because you can't get any services on the network without presenting your authenticated identity. So it both increases security and simplifies the access of the system services, and we see that as a trend, generally, across the IT market these days.

So the technology is sort of in continuous development with our partners in the industry, including the credit card companies, the card manufacturers, the carriers. So there are new features that are in a constant investigation. But the problem is very mature. It's been around for four years now. It's deployed in close to 300 million cards. There are lots of applications out there. It's not likely to just quickly sort of flip and change course. So I think it's a pretty stable platform and I think it's proven pretty viable and effective for most of the market it's trying to serve.

J.Schwartz Just so you're clear, we provide the software platform that runs on the chips provided by others. JavaCard is the software environment. Yes, there is a software environment on these things. These are, what, 16-bit cards?

A. Brenner They're moving to 32-bit cards. They typically have on the order of 100-k of memory.

J.Schwartz So the card technology evolves asynchronously to the phone, and every time you buy a new phone, you're probably getting different and new card technology.

Register I was interested in the talk about the volume play amongst developers, especially. I speak to a number of people and you always hear from people there's no deep developers on a SPARC box or a workstation anymore; everyone is using like PC's or something. So that kind of… Sun, as far as the studio or the tool desk goes … aspect goes or it's got a long way to go to get that volume, that density amongst the developer community where it's obviously blackballed and now they've got market-leading tools, that kind of stuff.

How do you intend to address that because it seems … you're going to have a big volume play where you're going to give the customer everything in a box, but whether they use it or not is open for question?

J. Schwartz How many folks do you know who develop on it? None, because they all develop on an x86 PC. So where you develop and where you deploy, don't confuse them, because I think increasingly, in a network world, where was TiVo developed? Probably on a Windows PC, actually.

I think there's an enormous volume of applications, Java applications, which are developed on Windows and deployed on large-scale 64-bit SPARC systems, huge quantities, and they aren't going down, by the way. WebEx is going up, not down.

M. Bauhaus To be specific, Studio's multi-platform in the Sun ONE Web Services Platform Developer Edition I mentioned is multi-platform, so Linux, Windows, Solaris, you take your choice. Again, people are developing to that strata of the Java Web Services platform, not the operating systems, specifically.

J. Schwartz A more interesting question as well is how many developers are building to Windows? I don't see that growing. I see the number of developers growing who build content for the Net. Guess what? That's open and competitive. Good luck, Microsoft.

Just think about how profound a shift that it. Ten years ago people wrote to the Win 32 API's. They built Windows applications. Go poll your CIO's that you talk to. Ask them how thrilled they are to write to Windows today. They're not. It's gone. It's been broken. Will they build a .NET? Maybe, but only if they can prove interoperability.

Computerwire It kind of strikes me as interesting that, again, there are definite market leaders, away from the developer question itself, there are definite market leaders on certain tools, in application servers, just for example, and Sun's got a long way to go. I think just a few minutes ago you said, "Yes, sure, we came to raise the game."

How are you going to really address that because, again, there is definite entrenched market leaders there? And again, on paper, it seems like you're going to give X customer or whoever everything in a box. Again, whether they actually use everything and become a true Java Web Services convert - the question is do you think everyone is using everything in that box across the entire market?

J. Schwartz I found, in my travels, and I encourage you to ask the same questions, that customers are very comfortable with Solaris. With Solaris they want to build a consolidated platform on which they can run their Web Services application. They're very comfortable in 64-bit systems, for those that run at scale, and if they're running render farms, the odds are good they've moved to stack x86 systems, but they didn't need Solaris to do it. So will there be x86 systems that are built for operating Web Services infrastructure? Absolutely. We believe we're at the front end of that. I mean who else is in front of us in that?

Do we believe that there's going to be a resurgence in the workstation market and more people are going to start developing applications on SPARC stations? No, to be honest. I think we see increasing momentum around building networked applications that are built initially on a desktop that is not necessarily related to the server on which they're ultimately deployed.

M. Petry We'll just take a few more questions, and I want to check with the folks on the phone again. … holler the question in the back of the room.

Handelsblatt Handelsblatt, from Germany. I have a question. There's a big portion in the United States for installation of Wi-Fi networks, far more than in Europe at the moment. Do you think that will slow down the adoption of mobile phones for data services? We are also saying why bother with narrowband phones if we can have Wi-Fi networks?

J. Schwartz I'll give you my flip answer. It's really hard to make a phone call with this portable. I'm picking it up and sticking it to my ear and it doesn't work.

A. Brenner It's been a big area of focus, right; we're trying to enable wireless data services with existing ecosystems and food chains. We're hearing interesting comments, but some of the carriers, including the European carriers, say, again, "There's no free lunch in this current environment, so we're watching it. If this becomes real, we'll go in and we'll dominate." So a lot of people are being a little skeptical about what they see happening here.

As far as everything that's being written about this being the next revolution that puts the PPT's and the existing manufacturers and devices out of business. So we're tracking it; a lot of people are tracking it, but that's not a main area of focus right now.

It is interesting on the mobile enterprise market. PDA with the wireless connection is an interesting target for employee applications. So we are tracking it there, and you're going to see, coming from us in the next quarter or so, some interesting platform announcements about how we're going to address that from a J2ME technology point of view.

J. Schwartz The thing that I found really surprising is that the demand for PDA's has sloped off compared to the demand for mobile devices. Why? Because they begin to get to the point where they're PDA's. And if you can make a choice between running a device that's about the size of a carton of milk versus one that's about the size of a phone, you tend to defer to the one that's about the size of the phone because that's what you're used to carrying.

I've watched a lot of the rollout. I just keep wondering; how do you make money on that one? I think free bandwidth, "Sorry. I've seen that movie," and there are a lot of broadband providers who are no longer extent because they tried to make money on something that was inherently undifferentiated.

Open Sys. Advisor So much of what you say is the potential for the customer to capture value depends on two things: number one, doing the math and coming to some conclusions about cost, but also just being willing to conceptually let go of either the models we've been working with for the past ten years, in terms of what is the conventional wisdom about who's right and what everyone should buy.

I'm wondering; in the what I'm sure is a large group of software developers who are looking to make some money on the opportunity, do you see or, more simply, who do you see as the most promising vendors to watch in terms of synergy between a new desktop and a new device form factor? It's a very broad question.

I was really intrigued by what you said about logging in and authenticating yourself using your phone, and combining this with what you showed us with your Net stations and adding that to the opportunities. Maybe not so much in the US for white-collar productivity, but certainly around the world for the potential of Linux on the desktop, doesn't that open up a whole new productivity environment?

J. Schwartz Let me try to give you a brief answer. Anyone who has a network relationship wants to monitize it wherever you are. So if you have a relationship with your bank, the bank wants to ensure they're available to you on your phone. If you have a relationship with a telecommunications company, they want to make sure that they've got access to deliver value to you through your desktop.

Greg Papadopoulos, who is Sun's CTO, has this great quote, which is, "All of us in the computer industry were dead wrong when we thought the replacement for the personal computer was a network computer. It was the personal network." I think we are all moving towards a personalized network where you are in charge; you have access; you are the one to whom services are being delivered. So that's kind of a conceptional level.

Secondly, from an infrastructure level, I have been surprised - given the number of discussions I've had recently with PC OEM's - at their level of interest in creating relationships and dialogs between this diversity of consumer devices and their PC's. Dell has a very interesting business opportunity in connecting their mobile enterprise devices, consumer devices with their desktops. Why? Because they're in the business of serving consumers, not simply being a conduit for Microsoft.

That's true, I think, with all the PC manufacturers as well as all of the device manufacturers. You're going to begin to see the connectedness between and among these devices begin to emerge, whether it's through Blue Tooth or physical cables or through interactions and authentication over the Web, in ways I think you haven't seen. I expect to see that in the next 12 to 24 months; I think we all do. I don't know if you've got any more perspective.

A. Brenner I think that's exactly right. I think the slides that you presented at the analyst conference are right on. Connection is more important right now than new features. For the market, "How to combine value out of a computing devices that are on my network and how do I present that to the user in a personalized way?" I think that's what's on people's minds right now.

M. Petry With that, I'd like to end our Town Hall today.

Back to top