Sun gets fat on thin clients

That's the message from Sun Microsystems, one of several computing giants that tried to popularize a thin-client approach to business computing. Instead of contending with full-featured PCs, workers with modest computing demands would use minimalist boxes that connected to a central server, where all computing tasks would be executed.

The approach never caught on in any big way, partly because most workers need some extra level of functionality and partly because plunging PC prices undercut any cost advantages.

But Sun is ready to have another go at it with Madhatter, the code name for a project intended to outfit businesses with low-cost PCs running on Linux software that can be easily configured by IT administrators. The IT folks can shift as much computing responsibility to the server as they see fit, power down applications to restrict access to certain Web sites or remove unneeded functions.

Curtis Sasaki, Sun's newly appointed vice president of desktop solutions, said Madhatter boxes would initially appeal to CIOs, equipping call centers and other divisions where workers require only a basic set of PC functions. CNET News.com talked with Sasaki about the Madhatter strategy and plans for StarOffice and OpenOffice, Sun's office software packages that compete with Microsoft Office.

Q: Madhatter is a fairly new entity--can you tell us briefly what it's about?
A: We spent a lot of time recently talking to our key enterprise customers, trying to figure out exactly what keeps them up at night, what the issues are. We heard something pretty consistent, and it boiled down to three things.

One was the fact that they're looking for ways to reduce their overall IT infrastructure costs, which obviously includes the cost of deploying desktops and all the software that goes on the desktop. With Microsoft's "Software Assurance" (the company's controversial new licensing plan), that woke them up even more to software costs. In a lot of ways, I think that drove them to begin to look at alternatives that might also solve problems besides cost.

The second thing we heard was, 'Can you help us improve security?' It turns out a lot of enterprises today deploy the same desktop to everyone, which is not what a lot of CIOs want. Giving the same desktop to a call center person and to someone in finance is probably not the best use of resources. Giving more fine-grain control of applications, being able to turn off access to certain Web sites--those are critical to what CIOs are looking for.

And the third thing is they'd kind of like to move toward open software--open platforms, open-source code, open protocols. They know that will prevent them from being locked in to one vendor, which they're finding out isn't the best model.

In addressing those, we started out with StarOffice. Because of the cost savings and the interoperability with Microsoft Office, it really was the first office product that got their attention. Beyond that, they came back and said 'What about the rest of the desktop?'

We feel the time is right both on the technology side and the customer interest side to deliver a full desktop environment that interoperates within an existing infrastructure. We can connect to an Exchange server, our SunOne middleware stack...That's what Madhatter is about--an open platform, interoperability with Microsoft environments, a complete desktop solution.

What makes this different from previous attempts at popularizing the thin-client approach?
We were one of the companies promoting thin clients. We still have a thin-client product today, the SunRay. Madhatter is different because of a couple of pieces. SunRay you can kind of think about as a video card encased in a piece of plastic, rendering whatever it gets from the server. It's secure, it's safe, there's no hard drive in the box. That's fine for a certain class of users.

There's also a class of users who are more power users or are mobile...If you unplug a SunRay from the network, you've got a black screen. Madhatter will let you work offline. What we now have is a multipronged approach for different kinds of users. The SunRay approach is still great for some people, but Madhatter gives you the option of a fatter client.

So will Sun be selling Madhatter PCs, a software package or both?
We're looking at a multipronged distribution strategy. There's a lot of enterprises that want the full packaged solution, with the hardware, with servers, everything preconfigured, preinstalled. We have a great sales force to deal with that and we'll definitely pursue that end.

The other ways are through OEMs (original equipment manufacturers)--there are a lot of PC makers looking for other ways to make money. We're in discussions with a number of them.

Cost has always been the main argument for thin-client computing--is it harder to make that case now that Wal-Mart is selling $200 Linux PCs?
Not at all. I think the more of those that are sold to consumers, the better it is for everyone in the non-Microsoft camp. It's all Linux-based, just like our platform. Having different companies address different market segments is a good thing. I applaud Michael (Robertson) at Lindows for trying to tackle the consumer space.

But I don't see major enterprises buying their systems from Wal-Mart. They want an end-to-end solution, they want 24-7 help, they want on-site service.

How central is Java to Madhatter?
We think Java will become a much more dominant programming platform. Enterprises want the ability to run multiplatform applications, and the only way to create those is in Java. The ability to leverage different computing platforms, even different Linux flavors, is very important.

The value proposition we're adding is taking all these layers of software and really integrating it well into a full system. We've seen in the marketplace other attempts at Linux desktops...basically what they've done is take a lot of open-source software, throw it on a CD and expect it to be a complete desktop. What we've been trying to do is look at all the things required for a complete desktop, and Java is a big part of filling in those gaps.

Will Madhatter use the standard Gnome Linux interface, or will you come up with your own?
CIOs let us know they really don't want to spend a lot of money to retrain employees. That means we have to provide a user interface already familiar to them, which means staying close to a Windows look and feel. That doesn't mean everything is identical, but things should be familiar, so nobody has to spend time relearning the system.

StarOffice just claimed its first victory with a major PC maker. Where do you see the major opportunities for StarOffice? Is it retail, OEM or enterprise customers?
We've had a lot of success at retail...What we've focused our energies on recently is to work with OEMs and Linux distribution companies. We're bundled with Linux as far away as China, with its Red Flag software.

We've also focused on hardware OEMs--that's sort of our next venue. We made our first announcement with Sony, and there's more to come. We also think there are opportunities in the service provider space. The big thing is that having the compatibility with Office creates a whole wealth of opportunities.

StarOffice had a lot of popularity first on Windows. But once enterprises evaluate software alternatives, one of the first things they want is to not have to pay for a Windows license. What we're providing is a way for enterprises to move off the entire Windows platform.

You've talked a lot about Microsoft, but isn't Corel's WordPerfect the main competition for StarOffice?
No, I never view Corel as the competition, because its market share is pretty low. And we've been working with Corel in OASIS (a trade group formulating an open, XML-based format for office documents).

I don't view it really as competition. I view it as an alternative to Microsoft, and the more of that that happens, the better it is for everyone else.

Has the settlement of the Microsoft antitrust case helped promote interest in software alternatives?
I'm not sure if it was the antitrust case. I think it's more Microsoft licensing policies of late. That and the economic situation. In the good times, nobody really cared about the expense of desktops, the expense of manageability, the cost of applying patch after patch. We're in a different world now, where every penny counts for CIOs. I think that's probably more what's opened up the industry. It's more about security, cost and open standards.

What are your goals for StarOffice as far as market penetration?
Distribution to me is really critical. I think the combination of StarOffice and OpenOffice (the free, open-source version of the software) is just a great one-two punch. You have an open version for people to try, and that opens up the market that much more. And when they decide they need the support and other resources we offer, StarOffice is there.

We've already had 10 million downloads of OpenOffice, and I think it's going to keep growing. From my perspective, every version of OpenOffice used and every version of StarOffice sold is one less user for Microsoft Office.

Some of the analysts have predicted we'll be at 10 percent (of the office suite market) in 2004. That would be nice. And I think there's room for growth beyond that. But I'm still being pragmatic. I don't think we're going to get 50 percent market share. But if I can get 15 or 20 percent, I think that would be pretty nice and pretty achievable.

Microsoft has talked a lot about adding XML functions to the next version of Office. Is that part of Sun's plans for StarOffice?
Our default software map is XML today. It's not a future thing; we're supporting it today.

I think the thing Microsoft hasn't really stated is that its XML format includes some binary codings that it doesn't plan to make public, so it's not an open XML file format. I don't think that does the industry any good. The best thing for the industry, end users and developers is to use fully open XML, so we really encourage developers to use those capabilities.