Andreas Typaldos - interview

From LXF Wiki

Impure thoughts

Andreas Typaldos believes not all businesses are willing or able to have a perfectly pure Linux setup. If Linux ­and his company, Xandros ­ are to thrive, he tells Linux Format, they need to accept it's a mixed platform world.

You might know Xandros as the company that took on the Debian-based Corel Linux in 2001. Indeed, its engineers still produce desktop distros. While these are praised for their extreme user-friendliness, the company has been criticised for its lack of community focus and for keeping some of its own software closed source. But more recently the Xandros CEO, Andreas Typaldos, has been steering the company into enterprise waters, producing corporate distros, server products and interoperability tools for administrators trained in Windows. Typaldos, a technology veteran and founding investor of Ximian and CodeWeavers, makes no apologies for being user-oriented and focusing on intuitive graphical interfaces, "much to the chagrin of many a command-line geek," he says. It may not fit your idea of free software in its purest sense, but Typaldos argues such pragmatism is the best way to eventual Linux migration. Nick Veitch met him to find out more.

Linux Format: Have you drawn any conclusions about the deal between Novell and Microsoft?

Andreas Typaldos: I think part of it is obviously that Microsoft are not happy about Linux. On the other hand, part of it is damage control. If you're a customer ­

LXF: I guess they know that practically all of their customers on the server side have some sort of Linux.

AT: See, I've been around all the way from the proprietary platforms, when you had Wang, DEC, Data General, IBM Series 1, and then you went to open systems Unix and Linux ­and even the other platforms in other areas like the database side. And I believe the reality is that nobody can force a customer into a standard. At Goldman Sachs and other places that are customers from my other companies, you know for a fact that they have Red Hat in the data centre and Active Directory as their identity management policy. They say, "Look, they've got to work together: you can't force me into one or the other." So I think part of it [the alliance] is a recognition or admission from Microsoft that Linux is here to stay.

LXF: I just wonder: the IP question that sort of hangs over Linux. I can't see a day when it's ever going to go away.

AT: I think, realistically, that as the days go by and Linux starts becoming a standard, who is going to take all this stuff out? It's like what happened with Windows: after a while, it became a standard even if you had issues like monopoly and all that. Nobody can rip that out ­ the world economy would collapse. The more time goes by, the more it becomes a moot point.

And the other thing is, let's say that you're Microsoft. Are you going to go sue Goldman Sachs because they've stopped using Linux; because of an IP problem? You'd win the battle and lose the war. I think, frankly, there was more danger out of SCO and those kind of guys.

LXF: Because they had nothing else to lose?

AT: Exactly. I can't see Microsoft suing major customers, even if they are in fact trying to win a PR battle about IP. Because, like you said, they have something to lose: in the end, the customer gets hurt.

The reality of it is, it's a heterogeneous world, it's a mixed platform world, the customer is in charge, and anybody who disregards the customer beats up their own parent. Thirty, forty years' worth of proprietary versus open standards war has taught us that. Where is DEC today? Where is Wang today? Where is Data General? It's not going to happen. There's a place where a Novell server is better, and a place where Red Hat's server is better, a place where maybe Sun is still OK, and a place where Windows is. Customers have a business to run. You can't really ask them to rip things apart.

LXF: Is that philosophy where you think Xandros is different from Novell or Red Hat?

AT: I think the value that we distro guys have to offer is to really contribute to the community ­ make a product better or participate in the development of the product ­ and to provide value-add [sic] on top of that community's efforts so that the product, this platform can be adapted in the real world quickly.

So our view is recognising a very simple reality, which is that the world is heterogeneous, number one; and number two, that people should not be spending time connecting things or integrating things ­ they should be plug and play. You can't say, "All right, take Linux, but you'll need five engineers to install it and three engineers to figure out what to do if it breaks tomorrow." If you have Windows-type administrators, you should be able to use them. So our fundamental philosophy is coexistence, interoperability, integration and the utilisation of existing skill sets.

Take the SMB [small and medium-sized business] market. It is primarily a Windows Server market. Why is that? Well, because most of the solution providers have Windows skills rather than Linux skills. They can't afford to be certified by Red Hat or have engineers certified by Red Hat before they sell a server for $500. And they can't go to a customer and say, "By the way, you need Linux skill sets or Linux administrators." That has limited Linux adoption in the SMB market. You can adopt Linux in the data centre at Goldman Sachs or Citigroup or wherever, but there you're quite often replacing Unix ­ you have skill sets already.

Really, if people like Xandros want to help the adoption of Linux we have an obligation to give to the community not just enhancements to the kernel and open source products, but to create tools on top of that. If I get a bunch of programmers to add some features to the kernel, is that more valuable than creating some tools [so that] all of a sudden the whole world can use Linux?

LXF: That depends on what the kernel patches might do!

AT: My point is that we have an obligation to do both. It is not simply having the open source, packaging it and giving it to somebody.

LXF: How many developers do you have?

AT: I guess that we have almost 60 engineers in North America, a few in Europe and about 15 in India.

LXF: And Xandros is defiantly not just a desktop distro company any more?

AT: No, we've been able to leverage what we did on the desktop to create a server product for the SMB market [Xandros Server], where it can use existing Windows-oriented admin skills. For example, if you are a small or medium-sized business and you're making a change to one service on your server ­ let's say your firewall ­ you shouldn't need to know, like an engineer would, that if you make this change you'd better make a change over here, over there and ten other places for it to work. Even if you are an engineer, if you make a change in one spot and then your wife phones you and says, "Our kid got sick, come home" you've got to get up and leave, and meanwhile you've made a change in one spot and all of a sudden the whole thing is out of whack.

LXF: Do you think that is the major barrier for people adopting Linux ­ they're afraid that they don't have the skills?

AT: For sure. But I think that on the desktop, the fact that some people are making the products easier to use, more plug and play, means that users don't have to learn something new ­ there's no command line stuff. It also means that application providers can say, "I can put my applications on Linux, because people can buy more Linux machines now," and that in itself at some point pushes the PC manufacturers to image it. But it's really a Catch-22, a chicken and egg problem.

In the enterprise or with SMBs, I believe adoption is going to be helped by providing tools that essentially allow you to cut out the cost of management. I don't want to have to remember five things to get something done. I have to get this piece from here, that piece from there? No, make it simple for me. Give it to me in an intuitive way that's simple and business function-oriented. I don't need to know what's under the hood.

LXF: You just want a nice little control panel, so you see Email ­ yes, I want that; Sharing ­ yes, we'll have some of that; and a button for when things go wrong that says Fix Things.

AT: Right. You want to say, "Take this whole thing, put it in a black box and allow me, from one spot, to manage something regardless of whether it is on this server or that server." That's the real world. That's what you want as a user, as opposed to me telling you to remember certain tools to manage each different server. That's our idea: a single pane of glass management. This `black box' becomes an engine and then can run on Novell, Windows, Linux or even the browser and connect to your management tools [Typaldos is referring to Xandros BridgeWays, a cross-platform sysadmin suite].

LXF: Can I ask you one question about this? What happens when your Red Hat server has a security patch for something that changes the way BridgeWays works? How does your management tool work now?

AT: Well, how would you do it at the moment? You would go into the Red Hat Network website, take the patch, bring it down and distribute it. You'd have to also remember that you have Novell in another department and Xandros in a third, and some people are running Windows. Why can't you have one console like a captain's bridge ­ where you can say, "Go to Red Hat, get the patch and send it out to the Red Hat servers; then go to Novell to get that and send i out to the Novell servers; and the same for Windows"? You're still doing the same thing. What you're really doing is you're making software do the work. Right now the work is being done manually.

LXF: OK, I can see that.

AT: We believe that for us to make a contribution to the adoption of Linux, people want us to make their lives easier. They want choice, but not at the price of complexity. If you tell them, "You have your choice, but now you've got to have one more manual to worry about," they'll say, "You know what? I'll stick to what I have. I'll upgrade my Windows PC to Vista rather than worry about learning something new."

LXF: Do you think there are too many choices within Linux at the moment because there are so many distros? Would it be a simpler world for people to get their head around if there were just SUSE, Red Hat

AT: I think in the end, evolution has told us that at some point you have elimination and consolidation. Right now, IBM and HP and others, all they're pushing is Red Hat and Novell. It's very hard to become the third choice.

LXF: HP now supports Debian as well, though.

AT: Right. I think that if, in fact, you have what I was describing before [Typaldos' laissez-faire philosophy], you will allow that choice. The reason why choice gets minimised is that nobody can afford too much choice. How many times do you go to a store and say, "I wish there were two products to buy from instead of 22? I've got to figure out all the features, do a comparison... just tell me what to do." So I think the choice is going to be limited, if in fact the price you pay is complexity.

Our approach, we believe, supports more choice. But in the end, the reality is that it will be very hard for many companies to exist if they're not able to put a lot of R&D behind it. I mean, nobody is as rich as Red Hat or Novell. Most of the distros don't have major resources. In our case, we have $60 million of R&D behind us. But how many people can afford to go spend $60 million? To keep adding value takes a lot of money, and you cannot be the same as everybody else. I think that is going to create more and more consolidation and reduction of choices. In my view I think it is also going to focus more on added value.

LXF: What do you think about Ubuntu now that they have the Long-Term Support version aimed more at the server market than just the desktop?

AT: Clearly Ubuntu has been a phenomenon, and it has been good for the industry. But in the end it comes down to the fundamental question: if you are the head of trading at a major investment bank in London or New York, are you going to bet your job or your customers and compliance issues on a purely open source platform because your techies like Ubuntu more than OpenSUSE or Red Hat? I believe that in the end, people want stability, they want value-add on top of the open source, so ­

LXF: In some ways, though, that's the same problem that SUSE and Red Hat had initially. If you were in a big bank, would you risk your job on choosing this weird Linux nonsense when everyone else ­

AT: But there's a difference. Red Hat was the first to market, had a billion dollars in the bank and was able to throw resources at it. There is a company behind it. If you are the head of infrastructure at Deutsche Bank, you call up Matt Szulik [the CEO of Red Hat], and if he needs to send you an army of engineers, he'll send them to you. As opposed to some guy saying he's downloaded OpenSUSE, Xandros or Ubuntu, and now your mission-critical trading system runs on Ubuntu. You say, "Wait a minute, what happens if you leave tomorrow? I've got a Windows guy here ­ who do I call?"

LXF: Ubuntu has certainly done a lot for the desktop.

AT: Of course. Debian is still the most successful platform out there, and Ubuntu is probably pushing Debian to say, "OK, let's have more regular updates, and so on." But for me... When I say "Cisco" to you, what do you think? Cisco is a company that has been created on top of open systems ­TCP/IP. But what you think about is the value added on top of that. Or take Oracle. Here's a company that was created on top of a relational database paper that was published by IBM 30 years ago. But the value of Oracle is not how many changes they make to the fundamental relational database concept. That's why in the end, our approach is to help Linux be adopted by providing the value on top of it that makes adoption easier. We cannot really go out there and do the same thing that everybody else did. Red Hat did it, Novell did it, Ubuntu is doing it ­ "Here's the open source and we'll provide you with a support contract." And now you have Oracle.

LXF: Yes, I'm interested to see how Oracle's Unbreakable Linux plays out. Tell me, where do you think Xandros will be this time next year?

AT: We hope to be able to see the fruits of this vision: pushing greater adoption of Linux in the SMB space and then greater interoperability in the enterprise space. We have an operation in Europe right now and we're expanding that, we have an operation in Latin America. We want people to see us not just as a desktop company but as a desktop company that [extended] its skills on the desktop to the server, to the management layer, to the applications layer. Similar to how Microsoft started out as a desktop company and became a server and applications company. I want people to be thinking of us like, "Red Hat, Novell... Xandros" That's a key positioning thing, an image thing, but secondly we want people to be able to see organisations adopting our products. I believe if they do, we'll make a contribution by levelling the playing field in the market so people are not afraid of getting more complexity if they get Linux. So we take Linux to places where it has not typically gone before. It's like the Starship Enterprise... LXF