Andreas, or simply, Andy, Bechtolsheim holds the distinction of having employee badge #1 among the thousands of people who have worked at Sun Microsystems in its 25 years of existence. Perhaps best known as the inventor of the Stanford University Network (SUN) workstation, he is among the four founders of the company, along with Vinod Khosla, Bill Joy, and Scott McNealy. Although he left the company in 1995 to found Granite Systems, a Gigabit Ethernet startup, he returned to Sun in 2004 via the company's Kealia acquisition. Bechtolsheim also was a vice president at Cisco Systems while away from Sun. InfoWorld Editor at Large Paul Krill recently spoke with Bechtolsheim about Sun's past and present situation, with Bechtolsheim chiming in on about topics including open source and the Intel and SPARC chip architectures.
InfoWorld: First of all, was Bill Joy actually the fourth employee at Sun? Or wasn't he more like the 15th or 16th employee?
Bechtolsheim: I think his official badge number was 6 because there were two people we hired on day one, and he came [onboard approximately] the next week. But he was considered a founder.
InfoWorld: And you wore badge number 1, right?
Bechtolsheim: Well, that's because I worked on the workstation for three years before the company started, although I think Bill worked on Berkeley Unix longer than that. So maybe he should have been number 1. The way the company started was I had a previous business that was actually licensing the design [of] the Sun workstation to other companies in exchange for royalties. And I had a whole bunch of people set up making these boards. It dazzles me that none of these companies had any clue about the workstation business. They had all the pieces, but they didn't see the market opportunity, and I was kind of getting frustrated because I was really more the engineer, the hardware guy, than the businessman. So Khosla called me saying, "Let's just start a company. This was the obvious solution to the problem of how to bring this product to market.
InfoWorld: I know you've left Sun, you've come back, and I think there's been a couple exits and returns.
Bechtolsheim: No, no, no, there was just one. I left in 1995 to start a Gigabit Ethernet company called Granite, which Cisco acquired in 1996, and then I spent seven years at Cisco until the end of 2003. And then I left Cisco to start a company, which Sun acquired in 2004.
InfoWorld: What are you doing these days?
Bechtolsheim: Well, my official title here is chief architect of industry standard [products] and I'm basically responsible for the product definition of our industry-standard architectures [offerings].
InfoWorld: How do you view Sun after 25 years? The company has had its ups and downs, mostly ups except for the last few years. How do you view the company at this point and where do you think it's headed?
Bechtolsheim: One reason I left in 1995 is that one of my concerns at the time was that we should be really building an x86-based server, and I was getting worried that the cost performance of those systems was edging up with the SPARC architecture. Here I am 10 years later or 12 years later actually doing what I suggested the company should be doing earlier.... People keep asking was this like a prearranged marriage? But the honest truth was I had no idea I would be coming back to Sun. I was actually happy building this [sort of] little business, but when Sun decided to enter the industry-standard market, it became an opportunity to build a much more significant business here. So that's why I'm back here.
InfoWorld: Were you the founder of the SPARC architecture?
Bechtolsheim: What happened is Bill, even though he was really the software guy, became convinced in, I think it was 1984, that Motorola and the whole microprocessor architecture they were using, which was the 68000, was just not keeping up, or not taking advantage of what the technology could do. And after some interesting investigations, we concluded that we might as well design our own architecture, which was, for a small startup company, quite the bold thing to do. What was correct about this was it gave us a performance lead for many, many years, and then [we grew] the company to many, many billion dollars. So it was actually a very successful decision. Now, I was not involved in the original SPARC chip design. However, I was involved in the design of SPARCstation 1, which was really the first high-volume product around SPARC, which was a pizza-box workstation that was launched in 1989. And that was actually Sun's best-selling product for a while, and that was followed by the SPARCstation 2 and 10 and 20. And then I guess I got a little tired of doing all these workstations, so I decided there was a new opportunity in gigabit networking. But now I do see an opportunity back in the server space, and that's why I'm back here.
InfoWorld: Do you think SPARC has much of a future at this point?
Bechtolsheim: Well, it's the mainstream of the company. In fact, it's growing again. And the price performance and power efficiency of these multithreaded architectures is quite compelling. Now, having said that, it is Solaris-only, so it defines the Solaris market, whereas the industry-standard architecture of course runs Linux as well as Windows. From a volume perspective, the industry's architecture is way ahead. But we have a number of very loyal SPARC customers, and we don't see a reason to switch anytime soon, so we will keep investing in SPARC as long as that makes sense.
InfoWorld: Would you say that Sun has righted the ship? I remember quarter after quarter after quarter where Sun was just getting bigger and bigger. And then during dot.com bust, Sun, like everybody else, slipped. Would you say that you righted the ship with the open-sourcing of Solaris and Java and that sort of thing? And also, what is your take on the open-sourcing of Java? I don't know if you ever actually publicly stated it.
Bechtolsheim: I wasn't at Sun when the crash happened. From the outsider's view, Sun probably reacted a little too slowly to the change in business fundamentally. So the company would have been better off at the time to downsize more quickly. But that was six years ago, so that's neither here nor there. The open-sourcing of Solaris is actually going very well. There have been over 5 million downloads to date. We don't actually know what people are doing with all these downloads. It's kind of fascinating to see people in Pakistan and India and China and all over the world downloading this code, and something interesting is about to happen here, I would think, and the same about Java of course. So perhaps the company should have opened up these systems sooner, but nobody's complaining about the fact that it's open now. And it certainly has received very positive feedback from developers and other people who want to make either extensions or just understand how this stuff is working.
InfoWorld: When you say positive feedback from developers, you're talking about Solaris and Java, or just Solaris?
Bechtolsheim: Well, I want to touch on the Solaris side. Java, of course, runs on all kinds of platforms, such as Solaris. But I'm not with the part of the company that's working on all the Java things, so I can't really speak to that.... Open-source software has a compelling advantage because the potential customers know that there's a difference in openness with open software than with any kind of closed software. And more and more governments and companies decide that that's the right thing to do.
InfoWorld: So you're in favor of the open-sourcing of Java?
Bechtolsheim: That was definitely the right thing to do.... Keep in mind that when Sun started, Berkeley Unix was essentially the first open-source operating system because you would get a copy of the source code just by sending a $100 check to Berkeley, which covered the cost of making a tape and sending it to you, right? In some sense, we should have kept that tradition all along, so the fact that Solaris was not always as open as it is today was basically a sort of mistake.
InfoWorld: So can Sun still make money selling hardware and services even though the software is now basically free in many instances?
Bechtolsheim: It's no different than Linux, right? Even though Linux itself is free, there's a cost for support and so on, and Solaris support actually, as far as that, is less expensive than a Red Hat subscription. So we are in fact competing with the Linux model.
InfoWorld: What are you most proud of in the 25 years of Sun, and what is your biggest disappointment?
Bechtolsheim: [There are] all the good memories certainly [of] the early days with the first Sun workstation, the first SPARCstation, UltraSPARC, and so on. All of these were major milestones that really propelled the company forward. And my personal involvement was by far the greatest with the SPARCstation series of products. Actually, really all the workstation products, but the SPARCstation really broke new grounds for the company in terms of just being a high-volume platform. One of the disappointments was that we didn't take advantage of the x86 architecture earlier. We should have done that in the 1990s certainly, not just a few years ago. But by the same token, there's ongoing innovation here, and it's not too late to re-enter that market.
InfoWorld: I keep thinking back to something like Open Look versus Motif. Do you think Sun has been a rebel sort of company?
Bechtolsheim: No, I think that's the wrong way to look at it. If you want to innovate, you sometimes have to throw things out and see whether you can make them successful. Because there's no easy way in advance to predict how exactly it will play out. Open Look is an example where something was not successful, even though we tried, we tried to give it away to the industry, but at the time, there was this counterforce called the Open Software Foundation, which was really the IBM, HP, DEC club to prevent Sun from growing, and whatever [we] were doing, they would oppose and do it differently. Now the Open Software Foundation, of course, wasn't successful either. There's nothing that remains from that. But the unfortunate thing is that what we were telling these other companies at the time was that the problem was Microsoft, not Sun, and they just didn't believe it. And a few years later that was more evident, but a lot of energy was wasted during that period of Unix companies trying to slow each other down, which really wasn't beneficial to anyone. But again, this was an industry matter of fact that we could just observe that not everybody understood open systems at the time, but I think it's clear now.