The First Major Open Source Database
by Doc Searls <doc@ssc.com>
4-Jan-2000
Doc Searls interviews Inprise President and CEO Dale Fuller.
Go to the Inprise (aka the Inprise Borland) Web site. There you will witness two highly significant developments: 1) the Corel-like linuxification of yet another leading PC software company; and 2) news that the company has open-sourced InterBase 6, the latest version of its cross-platform relational database. The product will be open source in all its versions, including Linux, Windows NT and Solaris.
InterBase is a first-rank relational database with customers that include Motorola, Nokia, MCI, Northern Telecom, Bear Stearns, the Money Store, the US Army, NASA, the Philadelphia Stock Exchange, First National Bank of Chicago and Boeing. It's hard to find an unkind word about this product, which competes with Sybase and Microsoft SQL Server, among other big-time database applications. So opening the source is a highly auspicious move.
Linux Journal senior editor Doc Searls talked about the move with Dale Fuller, Inprise President & CEO. Fuller came to Inprise less than a year ago to turn around what was once the premier franchise in software development tools - and in many ways still is. What followed has been a steady corporate drift toward Linux and the open source community, which is a natural constituency for Delphi and other popular Inprise Borland tools.
With this announcement, the company seems to have made a millennial shift into what might become a full-fledged Linux company (or at least as full-fledged as one can be when, like Corel, it still sells goods and services for other platforms). It also gives to the Linux community a very serious tools for migrating corporate databases off other platforms and onto Linux, among other significant things.
Doc Searls: I recently heard that you have some Linux credentials of your own.
Dale Fuller: Back in 1996 at WhoWhere, I developed the single largest Web site in the world running on Linux: Angelfire.com. It's still running on Linux and is one of the most popular sites on the Web.
Doc Searls: So why did you go with Linux?
Dale Fuller:When we did Mail City--an earlier site--we did it entirely in Solaris. It was roughly equal in size to Angelfire, but Angelfire cost one tenth as much because we built it on Linux rather than Solaris.
Doc Searls: Today let's talk about what's going on with InterBase. Why did you go open source with it?
Dale Fuller: Until now, InterBase hasn't been a core asset at Inprise--at least as a revenue source. In that area it's been small. But it has huge value in so many other ways. This is a world class, bulletproof, fully-tested application that just about everyone needs. So to gain momentum in the marketplace we decided to open source the product. Open source is a word of mouth market, and that's what InterBase needs. And InterBase is what the open source development community needs, too, because InterBase not only fills a huge hole for enterprises, but gives them a way to migrate in a serious way from one platform to another - especially to Linux. There is an explosion of demand for Linux and we can serve that demand by providing a free and open database product that is already proven and tested over the last fifteen years.
Doc Searls: How big do you expect this to be?
Dale Fuller: We expect it to be huge. We have a gigantic following today, by pure market share standards. But what you want with open source is something that serves the whole community, not just a nice piece of it. It's a different game, and one to which this is both timely and well-suited.
Doc Searls: And the prospects were pretty bleak in the commercial marketplace.
Dale Fuller: Exactly. Having a great product isn't enough. You need to spend big marketing dollars just to begin driving awareness. We don't have that kind of money, and we're not interested in playing that game any more.
Doc Searls: Especially since word of mouth is free.
Dale Fuller: Exactly. The Linux market is still young, and our timing is still early, and the mechanisms for growth are extremely powerful. There is a huge need for what InterBase has already been doing for a long time, and the word of mouth in the open source development community should be very strong, very fast.
Doc Searls: And it will help pull through lots of other products.
Dale Fuller:Yes. That includes the Delphi product for Linux, the J Builder products for Linux. These will all be optimized to work very tightly with InterBase.
Doc Searls:What kind of license will you use? GPL? BSD? Mozilla? Something different?
Dale Fuller:We don't have a decision on that yet. But we do want to do right by the open source community. So let me just say at this point that we're thinking hard about it, and open to input.
Doc Searls: How does InterBase stack up against--or with--existing Linux database products, like MySQL?
Dale Fuller:It's complimentary. MySQL is nice as far as it goes, but InterBase is the only commercially developed, tested and deployed open sourced database product in the world. Its Java support is unbeatable. So is its performance. Even before we open-sourced it, no other RDBMS was easier to port between platforms. And no other database could scale from desktops to huge, industrial-strength systems. Those advantages will only increase.
Doc Searls: Are you going to manage this out of your current corporate infrastructure, or is this too different?
Dale Fuller: We are setting up a separate company to manage the whole process. We'll try to follow the paths already beaten by Red Hat, VA Linux, TurboLinux and others. We want to consolidate a core development team, make our money on service and support and adjunct products, do the compatibility testing, and work with the community to make sure there is one source that doesn't fork.
Doc Searls: Do you expect that database - the InterBase database - can be part of a standard Linux distribution?
Dale Fuller: I do. I absolutely do. We're talking about building material here. That's the great appeal of Linux: it's better building material, and open source is a better building method. To run with that metaphor, you need more than just nails. You need boards. You need lumber. In most enterprises, that's the data. And in most cases the data is more important than the applications. You need something that's bulletproof, that doesn't have memory leaks, that's reliable and proven. You get that with InterBase. It's a huge advantage.
Doc Searls: What's to stop others from offering the same thing?
Dale Fuller: Nothing. But they will not be as mature. Here's a product that is already on sixty different platforms, used by hundreds and hundreds of thousands of people. Our 911 system in the United States is run on it. Now it becomes available to everyone. Think about what that means for enterprises that are only beginning to adopt Linux now. It knocks down a huge barrier to adoption.
Doc Searls: You see it as a way for companies to migrate to Linux?
Dale Fuller: I see it in a big way. We built this as a cross-platform product. Any data that you have on Solaris will immediately run on Linux. You just port your application over. You don't have to worry about your data structures. That's fundamental to what makes InterBase such a sweet product. It runs on Solaris, Linux, Windows, AIX, multiple flavors of UNIX, even Wang systems from years gone by. So when you're an in-house company looking to expand and get out into the world of application server technology, you need a way for your data to move with you. The good news is, with InterBase, you can do this easily and quickly.
Doc Searls: Instead of building over again on another platform, you move whole buildings.
Dale Fuller: You move CAD drawings from one output to another output, rather than having to FedEx or fax them. You move the entire architecture over.
Doc Searls: Because the data structures remain intact.
Dale Fuller: Exactly. And because we have the fifteen years of experience at doing it successfully across all these platforms, we bring a lot to the open source table. Including big-time customers who need to find a way to make the leap to Linux. There is something for everybody here. We want to see people involved in testing it, improving it, submitting their findings and changes back to the organization, and circulating back out the constantly improved versions that anybody can use.
Doc Searls: And you want to build an organization customers can buy support from.
Dale Fuller: Yes. That's where we make our money.
Doc Searls: I think you're among the first in the open source world to directly go after what we might call Practical IT. These are guys getting work done in large organizations, busily adopting Linux because they like its virtues. They are part of the movement in a practical sense, but not part of The Cause. Most of them I would guess are not heavy Slashdot readers, or if they are they don't post there. Which makes them a bit invisible. But we know they are out there, because we know from IBM, HP and many other large companies tell us that their companies are full of these people, who are making the choice to migrate enthusiastically from NT and various UNIX flavors to Linux. It seems to me that's your constituency.
Dale Fuller: Yes. And they care about the future, too. Take a topic like EJB, Enterprise Java Beans. Those beans become transferable; I can sell, give, buy or take templates for, say, spreadsheets. This constituency likes that kind of transferability across both vertical and horizontal platforms. Well, that's what open source InterBase will give them. No matter what platform you have, wherever you are, you don't have to worry about your data.
Doc Searls: So its virtues are a lot like Java.
Dale Fuller: Only not so closed. We're going open source all the way.
Doc Searls: And you think it will be easier to do that with a new company than with your old one.
Dale Fuller: Yes. What we're setting up will be the host organization outside our company. We don't want to do what Netscape tried to do, which wasn't very successful. We need a separate entity. We are funding it, but only as one source of funding. We want the community to help fund it, and own shares in it.
Doc Searls: Have you talked to some of the service players, like Linuxcare?
Dale Fuller: Yes we have. Also with Corel, TurboLinux, Red Hat and others.
Doc Searls: Are they ready to include you in their upcoming distributions?
Dale Fuller: We're moving to the next level of discussions. But I'm confident. This is the only serious database product in the world that is Linux compatible, that is actually available in open source form and that appeals at every level, in a very big way. There is no down side, for anybody.
Doc Searls: What does this do to Oracle?
Dale Fuller: It fires a bit of a cannonball across their bow. It is going to be much more difficult for them to follow along. Their whole game is about deployment and charging money after you develop on it. I think open sourcing InterBase will have a more immediate effect where Solaris and Windows NT are deployed. One reason is that it validates the Linux platform in a very significant way. It makes Linux far more competitive as a platform. With Linux you already have something that amazingly in a very short time is revolutionizing the world. This will step up that revolution by making databases a significant part of it.
Doc Searls: And in what situations does it threaten Solaris and NT?
Dale Fuller: Growing ones. NT is not the most scalable OS. Solaris is more scalable, but it's also expensive, so high cost itself is an impediment to growth. Customers are constantly looking for economically scalable solutions. The combination of Linux and InterBase gives them that. We also have a translator that transforms the entire data set out of SQL Server and into InterBase. And it will run under Linux or Solaris or any other platform. This is very appealing.
Doc Searls: The notion we've had for a long time is that you're building on an operating system platform. But in fact your building itself is mostly data, no matter what platform you build it on. Data is what you care about.
Dale Fuller: Yes. And what you get with an open source approach to data is far more safety. Because you can actually see what this database application is doing to your data. You can get under the floor and see what's causing the squeak.
Doc Searls: That's an interesting redefinition of data, because it presumes that responsibility belongs to the builders rather than the suppliers. It isn't like the old days when you wanted to trust the large company that would send out guys to fix your problem.
Dale Fuller: Enterprises increasingly want the flexibility to fix stuff themselves. Or have some choice about the help they get. You can still go out and hire somebody to fix your problem, but isn't it nice to know you have some options here?
Doc Searls: And that there is a wide-open market for those services.
Dale Fuller: I think what you're giving people is choices. Before they never had that.
Doc Searls: And you think scalability is the big issue.
Dale Fuller: I think scalability is going to be a gigantic problem, one that's going to make a lot of Linux converts out of a lot of big companies.
Doc Searls: One key is going to be exposing some of what's actually happening with Linux in enterprises. As we just said, that activity is not always visible.
Dale Fuller: In many cases, companies don't want to give away a technical advantage. It's a secret weapon. That's why I told my competitors at HotMail, "Man, we've got to stay Solaris all the time." Meanwhile we were marching down the path of Linux, making money because Linux was cheap.
Doc Searls: So you didn't talk up Linux when you were doing Angelfire.
Dale Fuler: No, we didn't. Another reason was that, frankly, we didn't hate Microsoft. In fact, Microsoft was willing to pay us millions of dollars to port one of our applications to NT. Maybe that put us out of the Linux mainstream at that time; but Linux has grown since then, and it now includes a lot of companies like ours, that use both platforms - and others as well. Companies like this take very pragmatic views of these matters. Basically, they want something that's bulletproof and scales. InterBase always gave them that, but now it will be in a new and better way.
Doc Searls: So you're clearly in the Linux movement now.
Dale Fuller: Back then we were deeply in the movement, but just not in a highly visible way. Now we're changing that.
Doc Searls: How does it feel to be a Linux company?
Dale Fuller: Great. I love what's happening now, because I was in the UNIX world a long time ago, in the early 80s, when AI was the rage. I worked on Common LISP at TI. The Micro Explorer was my project. The problem was, we had to do the sexy thing. We had to solve all the problems of the world, and that just wasn't a commercially workable goal. Industry kept coming back to us and saying, "Just solve one friggin' problem." We couldn't. Academia went for perfection and destroyed AI, because they didn't focus on the business aspects of it. I'm not seeing this with Linux. This time there is a consensus about the general direction of things, and solving business problems is a big part of it.
Doc Searls: As an ex-Apple guy, are you glad to see the company make open source moves with OS-X?
Dale Fuller: I think it's absolutely the right move. It will drive so much more adoption, and bringing in more ingenuity, more creativity. The more open you can make it, the more opportunities you have for developers to get in and actually make things worthwhile. That's the great thing about Linux. Nobody owns it, so it becomes a very creative platform. It makes me encourage Microsoft to open their OS.
Doc Searls: I don't think it's out of the question.
Dale Fuller: It's two years away.
Doc Searls: They need more of a sex change before they'll do it; but fundamentally they are a practical company, and they have no religion about keeping alive their dying products, which is a real advantage.
Dale Fuller: Meanwhile, there's lots to do here, and we're ready. And we invite Linux Journal readers to jump in and help with this thing, any way they can.
Copyright © 1999 Specialized Systems Consultants, Inc.
|
