|PHOTOS © GARY WAGNER|
Since its rather humble beginnings as a husband-and-wife freeware operation, Cliff Miller’s TurboLinux, Inc. has grown into a significant force in Linux — with over 150 employees and a strong hold on the Japanese and emerging Asian Linux markets. But TurboLinux, which used to be known as Pacific Hi-Tech, has been ramping up its North American operations of late, opening up new offices in the Bay Area, launching a North American reseller program, and inking a bundling deal with router vendor Linksys. Miller met with Linux Magazine Publisher Adam Goodman and Executive Editor Robert McMillan recently to talk Linux, and to explain why he thinks his company can succeed by focusing on something that many Linux companies don’t like to talk about: selling free software.
Linux Magazine: How did TurboLinux begin?
Cliff Miller: In 1992, my wife Iris and I started this Internet CD-ROM company in our basement. I was a graduate student at the University of Utah in computer science. I was in the Ph.D. program there and my wife was working at a company called Sirius that was later acquired by Novell. When Novell acquired them, they laid off most of the sales force.
At that time, we kind of looked at ourselves and said, “Gee. If these guys can run a business, why can’t we? We ought to just give it a shot.” So we decided to focus on taking U.S. software and selling it, first of all, in Japan and then in the rest of Asia.
Very soon after we started the company we began to get involved in the CD-ROM freeware business, and quite soon after that we began using Linux internally to power our Internet servers. So by 1993, we were also selling some Linux products.
LM: Where did the Japanese connection come from?
CM: There are two reasons for the Japanese connection. One is the business opportunity. The other one is really personal. I went to junior high in Japan. I was plopped into a Japanese junior high school — I stayed with a Japanese family for two years, and I was the only foreign kid in a public Japanese junior high school for probably 500 kilometers around.
LM: How was it that you came to go to school there?
CM: They have a saying in Japanese that roughly translates as, “Take your cute little kid and make him travel.” It implies that if you put your child through hardship, it will give him a special education. So that was really my mother’s goal, to get me interested in some other culture and language. And it really did that. I was the tallest in the school (including the teachers). I was almost six feet tall at 14.
Waiting for Janet
LM: So how big is Linux in Japan?
CM: Well it’s a pretty good size. We’ve distributed probably three million CDs. I’m not saying they’re all installed but we’ve distributed that many and when we bundled TurboLinux with a magazine called PC Work (which is not a Linux hacker magazine by any means. It’s sort of a mid-level, business-manager type of readership). Their volume for that month, I believe, increased by 10 percent.
And they said in the past when they’d done these special things, the best increase they’d ever seen was five percent. So the fact that there was a Linux CD seemed to carry some weight. People were really interested. We have a lot of registered users so we know it’s very real there. I haven’t seen any hard estimates on how many people use it, but my guess is it’s close to a million now.
LM: Has Linux penetrated as much in Japan as in the United States?
CM: That’s a little hard to measure, but it probably has. The thing about Japan that’s really interesting is when something catches on in Japan, it tends to really blossom. So I expect that a number of companies in Japan are looking at what their neighbors are doing. If they see reference sites, that’s important for them. Even more important is if they see a competitor deploying TurboLinux or Linux, then they’re going to start doing it too. They don’t want to be left behind.
You saw Apple. When they went into Japan, they barely sold anything in the beginning. Then they had this Janet Jackson tour and all of a sudden Apple was cool and then everybody wanted to use Apple. I think their market share went up to close to 20 percent in Japan — double what it was in the United States at its height. So when something goes well in Japan it really takes off. I think TurboLinux is really poised to take off in that market.
LM: Has the Janet Jackson tour happened yet, metaphorically?
CM: For Linux? Not yet. No.
LM: Are there any famous Japanese hackers?
CM: Oh yeah. You bet. The funny thing about this is people in Japan really are rather modest. Even though we have top programmers in our company, they really don’t want us to promote them as stars. So I can’t name names here, but there are people in Japan who have developed very important programs for Linux.
For example, the console program that will support Japanese, Korean, and Chinese characters was developed by a Japanese individual who is very well-known in Japan. There are other Japanese developers who have done a lot toward internationalization and localization.
Font programs is another area. Dealing with rendering and printing of fonts is something that they’re very strong in. So there are very good developers in Japan that we work with, and actually some of them work at our company. I’d very much like to promote them, but unfortunately, because of their modesty, they’re kind of anonymous.
LM: It’s interesting to think that though many of the Linux hackers and Linux users come from all over the world, Linux seems united culturally. Germans and Dutch and American and Canadian and Mexicans will read Slashdot and are sort of unified in a culture. You don’t see a lot of postings from Japanese though. Is there a Japanese Slashdot?
CM: They are participating quietly. Unfortunately, Asian developers tend to be more the listening crowd on those venues. Part of it is that English is not their native language.
We Sell Software
LM: What’s the difference between your business model and Red Hat’s?
CM: It’s a little hard for me to pin down Red Hat, but looking at what they’ve said over the years, it’s changed a fair bit. They’re saying that they’re going to turn into a services-and-support company, but when you look at their SEC filings, how much of their revenue was generated from support and services? I don’t think it was very much. So I think that’s their model. They also acquired a Web company and started this kind of dotcom type thing, and that may be a part of their strategy. When you look at their public statements for the SEC, you see a lot of mention about this Web and Internet play.
TurboLinux, on the other hand, has a business model that is maybe not quite as sexy, but it’s a tried-and-true method of generating revenue and building a business, and it’s selling software: We sell software. We take open source, TurboLinux operating software, we integrate commercial applications — in general they’re from third parties like IBM and so on — and we produce packages that are value-add so we’re able to generate margins on software products.
That’s very different from the other companies, and it’s a much more traditional approach, but I think we’ve seen companies build significant businesses based on selling software.
LM: We asked Bob Young when Red Hat was going to start selling proprietary software, and he said if they did something like that, they would be limiting their options. I’m sure you have heard his stump speech about creating this massive market — flanking Microsoft — by making computers affordable to the millions of people who can’t pay the licensing fees for commercial software. So is TurboLinux limiting its options?
CM: Well, that may be an argument for the desktop market, but we’re focused on the enterprise market. So if Red Hat is going after the desktop, TurboLinux is focused more on the enterprise, and that may be a very good distinction to make.
LM: Bob Young wouldn’t say that Red Hat is just going after the desktop.
CM: But you’re saying software fees are a significant factor. And if you’re talking about Windows’ fee of a hundred dollars or two hundred dollars, that’s nothing compared to enterprise hardware costs, for example.
LM: But enterprise customers want open source code so they can fix the software themselves.
CM: We’re open source as well, and we provide source code for TurboLinux. Just because our philosophy is open source, we don’t impose that on our partners. If IBM wants to have proprietary applications in source code, that’s up to them to decide.
I don’t think that IBM, Oracle, or many of the other enterprise software companies are tomorrow going to open source everything. So we work very well with them.
LM: So does TurboLinux plan to open source everything that it develops?
CM: We’re planning to always open source our fundamental operating system, TurboLinux. There may be other applications that we choose to release in a different manner under different licensing models. But the fact is that we’re targeting the enterprise market and we do integrate our open source TurboLinux operating system with third party commercial software packages to produce something that is easier to install and performs better than if you take all the parts and assemble it yourself.
LM: Is your clustering software open source right now?
CM: The clustering software is composed of a few different parts. There’s the kernel patch, which is GPLed; the management and installation programs that are GPLed; and the cluster daemon that resides on each node, monitoring whether those nodes and the accompanying nodes are alive or not. We’re staggering the release of that software. So initially it’s closed-source, and then six months after its release, we release that source code.
LM: What impact did Red Hat’s IPO have on you guys?
CM: It occurred on my birthday, August 11, so I remember the date. I think it was generally a good thing. It brought awareness to the market and showed that Linux companies were part of mainstream computing. So in that sense I think it was very good.
LM: The IPO gave Red Hat a lot of momentum. Was there a sense that: “Red Hat is out of the gate and now we have to watch out?”
CM: Well, look at CompuServe and AOL. Who was the leader six years ago? Certainly CompuServe. And then boy, out of the blue comes AOL. So it’s not always the first ones who win.
Robert McMillan is executive editor for Linux Magazine. He can be reached at email@example.com.