Sun exec battles "phantom NT"

The president of Sun's Solaris software efforts makes it clear that if there's going to be an alternative to NT, it will come from Sun.

6 min read
MENLO PARK, California--What's a company to do?

First discussed in 1996, Microsoft's Windows NT 5.0 Workstation and Server operating system still has not seen the light of day as a shipping product. Yet competitors like Unix-on-RISC chips server operating systems leader Sun Microsystems must keep the Microsoft upgrade squarely in its sights.

News.com sat down with John McFarlane, the chief of Sun's operating system efforts, to gauge the effects of NT 5.0's promises, get McFarlane's take on the state of Unix and freeware, and learn the role of Java in Sun's Solaris software plans.

The latest upgrade to Solaris is scheduled for release this October, an update--currently called version 2.7--that will complete the software's road map to 64-bits, a technology that allows users to essentially handle larger chunks of data on their machines and will also include new clustering software functionality.

Through McFarlane's glasses, NT is currently deemed a low-end nuisance, which could--in the long run--be a dangerous threat, a posture that could be deemed arrogant given the unquenchable thirsts of the Redmondians. But the former Northern Telecom veteran who ascended to the president of Sun's Solaris software efforts this spring makes it clear that if there's going to be an alternative to NT, it will come from Sun.

News.com: What are your thoughts on battling NT 5.0, which seems destined to ship in mid-1999 at the earliest?
McFarlane: We've been battling NT 5.0 for about three years. Microsoft comes up with all these charts and hype. It's really hard to battle a phantom--there's some figure in the mist with some alleged functionality. It's delayed again. I hear it's closing in on 40 million lines of code. Solaris is 11 million lines of code: very tight, very well structured, highly reliable. God help me, I don't know how they're going to make 40 million lines of code work when 20 million-plus of it is new vs. 4.0. They keep saying that the enterprise offering is 5.0--it's got an unpredicted delivery date, it's got bags of code, I don't think they have a clue how to test it for enterprise-class performance. I hate to win because my competition loses, but NT is in deep trouble and I would not want to be the product manager--I would be looking for a place to hide.

What is your view on the NT project itself?
They've done some good stuff on the commercial desktop in terms of integrating the Excels and the Words and so on. They've kind of taken their commercial desktop design center and said: "Gee, let's do everything." It's brain dead, the surest way to fail I know is to try and take all these complex, high-scale, high-reliability, demanding applications and pretend it's like integrating Excel and Word. I think where Sun excels is our design center has always been the Internet. It baffles me how na?ve Microsoft could be about the complexity of what they're trying to build.

At what level do you respect their market power and ability to sell software?
As a company, they're very aggressive, great marketing, fantastic drive, and so on. And they're great copiers. They're ruthless. And they have a lot of money to throw at marketing and a lot of monopoly leverage on the desktop to arm-twist and crow bar partners and channels.

Have you seen pressure on traditional customers of Sun to use NT more and more?
The pressure that I'm hearing from our customers is largely from those corners of the organization that have Windows on the desktop and project that experience into the back office or project that upstream into the server market. The [management information systems] directors and the [chief information officers] see a mixed network and don't want to have NT anywhere near mission-critical or near a data center, but are getting pressure from the line-of-business owners and the departmental people that are using Windows on the desktop. That's leveraging NT--not from a technology perspective, but from an association perspective.

Does Sun or Microsoft have to win for the other to lose?
The market for servers in the year 2000 or 2001 is around $55 to $65 billion, so it's a huge market. If it's only NT and Solaris servers, I'd gladly take 50 percent of that market.

Is there a danger that Java could become as fragmented as Unix?
I think the straight-forward answer is that we have 100s of licensees, hundreds of thousands of developers, and God knows how many [independent software vendors] and we've only had two minor incidents: One with Hewlett-Packard that we're working out and, of course, Microsoft, which is hell-bent on trying to block, so I don't see any balkanization. People all recognize the value of standards and all recognize the value of writing the application and having it run in any application environment.

How do you view the state of the Unix market?
I may sound a little aggressive but I think my key competitors now are NT and their intent to encroach on my space and the data center-MVS and System 390 [from IBM]. If I talk about key competitors we're targeting in terms of our investments and where we're going, those are two of the major competitors we're targeting. HP is in decline, Silicon Graphics is in decline, Sequent Computer Systems--these companies are either losing money or losing business. The only other Unix that had any growth--everybody else was negative--was Linux. And go Linux. A Linux desktop is not a Windows desktop. I see the HPs and the IBMs converting to systems integrators and service providers. They won't care--they're getting out of being in the products business.

What role can the freeware movement play in the growth of Microsoft alternatives?
I think it's really an interesting trend. I think it's a really powerful and valuable software movement, particularly for innovation. I personally wouldn't want to live in a world where the only desktop was Windows and everything was Windows--I think it would be quite boring. For us, these are our friends in the battle to maintain open standards and sustain innovation in the software industry. We want to make more of our stuff freely available and we just want to see it keep growing. The compelling value is the huge amount of innovation we can tap into around the world that isn't smothered because it's proprietary.

What sort of challenges do you face in addressing the perception that Sun is a systems company, with software relegated to an add-on status?
There's clearly a lot of end-user interest. Interestingly now, more and more customers are asking for Solaris servers. From both resellers and [value-added resellers] we're getting feedback. We're getting pull now around the Solaris brand. Certainly, there's more work to do. When you look at the marketing investments Microsoft makes, we're in a constant state of paranoia about getting enough branding information out there.

How does Solaris fit into the movement to build Java applications on the server?
I think it opens up a huge new opportunity for the traditional Solaris business because Unix interfaces and the numbers of new developers developing to Unix is probably flat or in decline, roughly. That doesn't mean applications aren't growing. The volume is still driving ahead, but a lot of the new developers are going either to Java or to the Windows interface, so this opens up a huge new realm of application opportunities. Now many [independent software vendors] are coming on board that we can now partner with who may not have written to the Unix interfaces but can now write to the Java interfaces and yet we can partner with them the way we've partnered in the Unix space. It just expands the many [independent software vendor] partnerships we can do. We've got 12,000 applications that run on Solaris today and we will now be able to leverage the 700,000 to 1 million Java developers in addition.

Is Sun's suit against Microsoft due, in part, to the possibility that Microsoft could fracture the server-side Java effort?
Certainly they dominate the commercial desktop, but I don't think anyone wants nor needs a Microsoft monopoly on the server side, and there's a huge amount of momentum around Java, particularly Enterprise Java Beans.