LWM's editor-in-chief, Kevin Bedell, spoke with Sun's John Fowler about what's new - with Sun, with Fowler's new position, and with Sun's place in the open source world - and learned a little history as well.

LWM: Thanks for talking with us John, and congratulations on your recent promotion. Can you tell us a little bit about your new responsibilities?
John Fowler: As you know, I've been at Sun for about 14 years now. Over the last two years, I was the chief technology officer for software reporting to Jonathan Schwartz. Jonathan decided to reorganize the systems teams; prior to this we had two systems teams, one doing exclusively SPARC/Solaris for very large computer systems, as I like to say computer systems for which you need a forklift; and then we had another group doing both SPARC and x86 systems called Volume Systems Products. What Jonathan decided to do was to take some networking technology that we had, as well as the x86 systems team, and make that a direct report to him as a small group with the intent of growing that business rapidly. That's what I'm heading up, which is quite a different role for me, but an exciting one.

LWM: I understand you just got back from China. What are your impressions of what's happening with Linux there?
Fowler: We have a major event where once a quarter we collect a lot of our announcements and have a customer seminar that includes technical seminars and education as well as product announcements and activities. We held it in Shanghai for the first time this year, and there were actually more than 5,000 customer representatives there from Australia, Singapore, Asia Pac, Japan, China...basically that whole region. First of all, desktop Linux is of extreme interest across all the Asia Pac rim -no doubt about it, they're extremely interested. A lot of the discussions that I was involved in revolved around that. It's a vibrant community, a community that's growing; there's a lot of prosperity over there and they're looking to grow and use Linux in a lot of aspects of their business.

LWM: Did you have any specific announcements that you were there to talk about?
Fowler: In terms of me and my new role, I had a couple of small things to announce. Sun, of course, announced all sorts of things in both software and hardware, including in our identity product, the Java ES system, actually on Linux, which is our middleware stack; a new agreement with Fujitsu on hardware; etc. In terms of my own product, I only announced a couple of small things - one of them in the packaging of large-scale grids, of my 2-way server product for people doing primarily high-performance computing and numerical analysis. Previously I had that with Xeon-based systems, and now I offer it with Opteron-based servers. We showed some things; we have a workstation coming and some other server products, and we actually showed them at the show but we didn't formally announce them. The reason for that is we're holding off announcing until we're ready for true volume manufacturing. One of the things we discovered is that the demand for the Opteron products is very high, and we actually have a lot of volume.

LWM: It seems like HP, with their recent agreements with JBoss and MySQL, is working to deliver an application stack on top of their servers. Do you think customers are moving away from the build-it-yourself approach to application platforms and looking to buy a complete solution now?
Fowler: The Java ES system, which is something that we announced a year ago and have been shipping for a while now, is the repackaging of our own middleware, which is Web server, application server, directory server, portal identity management, and management clustering. It's fully integrated and works on SPARC, Solaris, x86, as well as Red Hat Linux. So our general belief is, absolutely, customers will in many cases want to make some best-of-breed choices, but the middleware stack more and more is becoming something where you want to choose a whole bunch of capability and be sure that it works together as opposed to doing all of the integration yourself. The driving force behind that is simply cost. If the customer has to hire a whole bunch of engineers to do a whole bunch of engineering to make it work together, it costs more money and takes more time. They're looking to actually save time.

LWM: How does Linux fit into the plans for the x86 line of servers that Sun's coming out with now, and how would you position that compared to the Solaris x86?
Fowler: We qualify for and offer for sale Red Hat Linux and Novell SUSE LINUX, so those are the ones that we primarily work on. We include Solaris x86 with the systems and offer a lot of varieties for Solaris. From a positioning standpoint I'm happy to provide the OS that people want to solve their particular business needs. With Solaris I have opportunities on pricing, as well as all sorts of capabilities that are much more difficult to offer with the Linux vendors. So I'm being pragmatic about it in that we're here to solve a customer problem; if that includes Linux or even Windows, our systems can be used for that. Over the course of the year, we're going to be offering very attractive pricing and promotional offers with Solaris as well because I don't actually have to pass the cost on to another vendor. I can choose to pass the customer the savings. That's what we're going to do there-just be pragmatic about it.

LWM: Speaking of Solaris, Jonathan Schwartz announced recently that Sun was planning on releasing Solaris as open source; what exactly did Jonathan say?
Fowler: I actually wasn't at the press conference so I don't know what Jonathan said; a lot of people have paraphrased it to me though. I think of it as a general comment; we're always trying to create a larger market for developers and for volume, and in some cases source availability can help with that. The word open source means lots of different things to lots of different people. Solaris source has been available to a lot of different people for a long time - visit any university and you'll find Solaris source - so source availability has not been an issue. We're always looking at how to make a bigger market. Will additional tools, more programs, more availability create a bigger market? And we certainly we do that for Solaris as well. Beyond that, we're not actually announcing anything. Between GNOME, OpenOffice, StarOffice, Apache, etc., Sun has a long history of supporting the open source community, so chances of us providing more source over time are good.

LWM: I've seen a lot of situations in which Sun has contributed or given whole products to the open source community, for example the OpenOffice project.
Fowler: OpenOffice is a really good example because not only did we contribute OpenOffice, but we moved the source base forward. I think that there have been a tremendous amount of long-standing investments there in making that successful, which of course is of benefit to the entire community, whether they be Linux or whatever.

One of my prior roles at Sun was to manage the HotJava Browser -if you're interested in history, it was a browser written in Java. When Netscape actually spun off from AOL to form the Mozilla Foundation I canceled the HotJava Browser, which was our own product, and put the engineering team to work on Mozilla. In a similar timeframe we actually started working on GNOME, with the very early GNOME stuff, and declared our direction. We weren't doing CDE anymore, we were doing GNOME. So many of the actual activities in the open source community for the Linux marketplace- Mozilla, GNOME, and OpenOffice -actually started a very long time ago, five years plus at this point. We've had a longstanding contribution and engineering investment all the way through that period.

LWM: How do you think that pays off for Sun?
Fowler: I think there's a whole bunch of obvious ways. In the case of the desktop, and the reason we invested in Mozilla, GNOME, and OpenOffice, we believe that together with other people we can create a larger possible desktop market. Together with other people we can create better products, and I think those products in particular are the ones that are very well suited to that. So the benefits for us right now are very simple. You can see that embodied in the Java Desktop System for both Solaris and Linux. This part of the Java Desktop System actually comes from these open source efforts that we started years ago, so it's enabled us to actually have products. I think that those are good examples of success stories. Now there are lots and lots of open source projects that aren't successful. Sourceforge is a virtual minefield of of open source projects - but those in fact are ones that are good examples of success stories.

LWM: There's a book called The Innovator's Dilemma, and in the book the author, Clayton Christensen, describes all the reasons why big technology companies have trouble adapting when disruptive technologies come into the marketplace. What are your thoughts on that?
Fowler: I think it's generally true, whether you're an organization or a family, that if you get used to doing things in a certain way it always takes more effort to change than to keep doing things the same way. That's always the case. I think it's a bit of an exaggeration to claim that it's true of everyone. I think IBM is certainly an example of current time of a company that has done some adaptation. And we are going through a big process of adaptation now. I think that change is always hard, and you have to always be searching for the next disruptive technology. If you look at Sun's history, we started with workstations and we disrupted Apollo and a bunch of other guys with a different approach to workstations. And then we did it with mini-computers and servers where we disrupted, particularly, digital, but also other folks, with that whole technology. Then there were the bigger servers -that was traditionally strictly a mainframe technology. And also software, with Unix and open standards along this history, long before those folks were capable, and then with Java with portable layers of writing applications. So, what are we working on today in software? We're working very hard at leading in portions of the middleware space such as identity and Web services, which we think are extremely disruptive, and also in middleware pricing, middleware distribution, and there's a whole range of projects and technologies that are doing that. And then in hardware with networking with others. I think the trick for a company is to always be looking for the next thing, and be willing to go bring it to market regardless of what the issues might be.

I want to close by going back a little bit to what I'm working on - I'm working on actually growing the volume of x86 on systems in the marketplace, which obviously would be of interest to folks running Linux, but I'm also working on high-performance networking and the fabric that interconnects systems to build high-performance grids and stuff. So, you know, be on the lookout because we're going to have some interesting products in that space that Andy Bechtolsheim is designing, together with the software that goes with it. We're not planning to be boring this year, so stay tuned.