Interviewers: Scott Swigart and Sean Campbell
Interviewee: Bernard Golden
In this interview we talk with Bernard. In specific, we talk about:
- The emerging impact of open source as a paradigm shift
- Competition between open and closed source entities
- Geographic differences in the adoption of open source software
- Future directions for open source, including the integration of commercial motivations
- Maintaining customer loyalties with open source or proprietary software
Sean Campbell: What can you tell us about your background and Navica to get things rolling?
Bernard Golden: Navica does open source management consulting, which broadly means helping organizations figure out what they should do about using open source more effectively, as well as helping them address any related questions or risks that may arise.
My background is out of big IT shops and enterprise software vendors, and to be honest, I was fairly skeptical about open source when I was first exposed to it. I had sort of a forced opportunity to work with it, through working with a client who had spent all their money on a failed project.
They needed to do something without any sort of budget, so we were forced to turn to open source. I had a very, very positive experience with that, which left me with the impression, “This is going to transform the way IT gets done in the future. I want to understand that better.”
That experience set me off on a voyage of discovery about open source, where I started to understand its dynamics and the implications for large enterprises. I ended up writing a book on the topic called Succeeding with Open Source, which was published a few years ago by Addison Wesley, and through that, I ended up consulting to organizations as well.
But really, it was driven by my perspective that said, “This is a very powerful way of making software capability available to organizations.” And any time you have that kind of change in the dynamics and economics of doing business, there are profound effects.
To me, this is analogous to the rise of the PC, which transformed the economics of the way that information processing was done. Look at the impact it’s had on our economy and the way we live our lives. I believe that open source will have that kind of impact on the software infrastructure.
Scott Swigart: Let’s unpack that a little bit, in terms of a continuum like the one you mentioned with the PC. That unfolded from mainframes, green screen terminals, and so forth, with the PC creating a paradigm shift for industry and consumers. You suggest that open source has a similar transformative power.
With the PC, we’re maybe 20 or 30 years down the path. Where do you feel like we are on the open source continuum, in terms of how far into the transformation we are? And looking into your crystal ball, when we get further along, what do you think the landscape will look like?
Bernard: My sense is that we’re pretty early into the open source transition and its full impact, and if you’ll recall the early years of the PC, it mostly had resistance from within organizations, particularly from IT organizations.
PCs were viewed as being functionally deficient compared to mainframes or mini computers. They were sort of toys that did this, that, and the other thing. Serious organizations mostly looked at them in comparison to what they already had and saw big differences, which they interpreted as negative.
I think we’re still at that stage with open source. You still hear lots of discussions from people around things like, “Well, how do people make money out of it?” That implies that they are reluctant to use it because it doesn’t fit the standard economic model they’re used to. You also hear concerns like, “Where do you turn for support?” Again, they are accustomed to getting support in a certain way, and they are skeptical of an unfamiliar model.
All of that suggests that we are still at the “it’s different and therefore not as good” stage, in many respects. That’s a time of powerful tension, because what causes the thinking to overcome that resistance is the fact that it brings incredible power and control and economic benefits. And that’s very analogous to the PC.
As I said, there was a lot of resistance to PCs, but the benefits that they brought were so undeniable that they just kept chipping away at that resistance and eventually overwhelmed it.
I believe that we’re still fairly early in that move within open source: certainly further along than we were a few years ago, but not anywhere near where it’s going to end up.
Scott: One thing about comparing source versus closed source is that it’s really an asymmetric comparison. People want to think of open source in terms of features, functions, roadmaps, and support and software licensing costs, and it just doesn’t fit in that box.
I think there’s a good analogy there, using your comparison with PCs versus mainframes and the big iron UNIX systems of the day. That was an asymmetric comparison as well. People looked at the PC and said, “Well, yeah, but it doesn’t do all these things.” It took people a while to realize that the new model also did new things that the old model couldn’t, and over time, those new things gradually became more important.
What do you think are some of the asymmetric strengths of open source–some of the things that people will look at five years from now and see as key strengths that eclipsed some of those traditional enterprise software strengths?
Bernard: I think people will start to see that there are real strengths in terms of quality and support. It’s different, but in some respects, it’s much more satisfactory. I think that people will start to sort out the question of costs and recognize that if you figure out a way to take advantage of open source, the cost structure can be very, very attractive in many areas.
I don’t think we’ve arrived there yet today, but I believe that the economic characteristics of open source will really change the ways that you can use software.
Once people start, for example, comparing the use of one copy of MySQL to one copy of Oracle, they start to see the possibility to build novel kinds of applications when they’re not constrained by budget structures or growth patterns or things like that.
They will find that they can really innovate more freely. That’s extremely powerful, because that’s a place where the strength of open source starts to come forward in a way that traditional, proprietary, licensed software cannot really respond to.
Sean: Consider for a moment the case of a closed source company that feels they just can’t open source their product. For example, maybe they feel like the IP is way too valuable to do that, or there are any number of other reasons it may just not fit within their business model.
What are some appropriate strategies that come to mind that such a company could take to counter the strengths of open source as you’ve laid them out, outside of open sourcing their product?
Bernard: They could find ways to reduce their disadvantages relative to open source by doing things like speeding up their release cycles to match the release cycles of open source, or delivering better quality technical support.
Beyond that, they could figure out ways to leverage the fact that there are people out there who will use their stuff early on and provide lots of quality feedback around it, in a way that fits with various open source models.
Finally, one thing that’s very clear is that they could embrace the places where their stuff works with open source and be much more aggressive about supporting open source, in terms of aligning and interoperating with it. Microsoft, for example, is making some pretty strong strides in that direction, recognizing that it helps them if they’re able to work within an ecosystem that is powerful on its own.
Sean: On a related note, you can’t help notice that people all over the open source world are creating a ton of buzz around the iPhone, even though it’s probably the most closed platform in existence. You can’t even push data to it, as of the announcement today, unless you push it through Apple to an app.
And yet, you hear huge acceptance of it, everywhere from OSCON to the fact that Raven Zachary left 451–or kind of transitioned away, depending on how you look at it–to go from being an open source analyst to somebody focused on the iPhone.
What do you attribute that kind of dynamic to?
Bernard: Well, I left out one blindingly obvious strategy, and I’m ashamed that I didn’t mention it. You asked how a company that doesn’t want to go to open source can respond, and the answer, first and foremost, is to build great products. Bring stuff that isn’t available through open source that’s so distinctively competent that people say, “It’s not as free as I’d like it to be, but it’s still really great.”
Apple products tend to be really great. They’re just really well done, easy to use, and their stuff really hangs together well and works well together.
Sean: It seems to me that, at some level, if the product’s great, nobody really cares if it’s open or closed. Of course, they may care down the line, as in, “Well, it was a great product, and I loved it, but I’m still locked in, and I still don’t like that idea.”
I mean, you can have this whole tactical conversation, but if somebody built a fantastic product, generally speaking, no one from consumers to enterprises really seems to care whether it’s open or not; they’ll just go buy it anyway. It’s when products are one step below that–maybe just good enough–that the other factor starts to come into play.
Bernard: I think if you come out with mediocre products that don’t really offer a significant benefit, it’s hard to make a claim that says, “But you ought to use our stuff rather than using that other stuff that’s got such economic advantages.”
Sean: Right. Let me switch to a different area. What initial observations have you made as you talked to people in other countries? What other countries did you talk to? What were some of the initial differences you noticed with open source adoption?
Bernard: From a national perspective, there are issues of sovereignty for organizations in other countries. We hear over and over again, “We want to paddle our own boat, and we’re uncomfortable relying on another company, an American company, no matter how benign they are, just because this is a key national capability for the 21st century. And we just don’t want that dictated by something located in another country.”
For the report, we ended up talking to Brazil, Korea, South Africa, France, and a number of other countries. And to a greater or lesser extent, sovereignty underlies many of those things, including developed nations, like France. They have a set of sovereignty issues that they choose to emphasize.
And then, beyond that, looking at developing nations, there are very strong economic reasons as well. In part, there’s a desire to inculcate a native capability in IT, and they feel that that’s easier with open source. So you’ve got countries that are trying to build up local IT talent, local IT companies, and they’ve pretty much figured out they won’t be able to do that with proprietary software, so how about open source? That’s a way that they can do that.
And of course, for many of them, it’s a pure economics issue, in that they’re not in an economic position to pay developed world prices, because they’re developing economies. I think that’s going to be a very powerful driver for open source throughout the world and a real challenge for proprietary software companies.
Scott: Way back when, I wrote hardware diagnostics on UNIX systems, and we were using the GCC compiler. Open source, obviously, has been around for a very long time.
And one of the things that’s really driving it today is the fact that Linux has reached a certain level of maturity. There is no longer just an open source tool chain to be used on some proprietary OS, but now there’s a very viable open source operating system. Companies like Red Hat and Novell and Ubuntu have wrapped themselves around making it a completely legitimate offering.
First, Linux started to make a lot of inroads into proprietary UNIX systems; the low hanging fruit was to move from proprietary UNIX to Linux, and a lot of organizations went that route. People kind of felt like, “Hey, open source is really coming into its own. This is may be the way software’s going to be built 20 years from now.”
Then some of these open source companies started getting bought, and initially, that was seen as great news. Citrix bought XenSource, and people trumpeted: “See, it proves the business model of open source. There’s money here, and real companies of value that are worth acquiring.” Sun Microsystems bought MySQL AB for a billion dollars, and again, people said, “See, this proves the open source model.”
At some point, though, there was a sort of uneasiness that started to set in, where people started to be concerned that these commercial entities weren’t going to be the best stewards of these projects, since even if they have the best intentions, they’re struggling to be profitable. And when Citrix bought XenSource, there was a lot of kind of consternation by contributors who said, “Well, you know, I really like Xen, but I don’t really want to view myself as being a coder for Citrix.”
And Red Hat shifted from Xen to KVM, partially, I think, because they viewed KVM as being a superior virtualization technology, but I think also partly because they just didn’t want their virtualization wagon hitched up to Citrix’s horse, and so KVM gave them a little bit of independence. And it’s supported by Qumranet, who also might be acquired.
Even take a look at Red Hat–I think people get caught up in the fact that Red Hat is such a standard bearer for open source. It does a lot of great things without compromising its ideals, and it’s 100 percent, pure play open source. But Red Hat could get acquired by somebody bigger. They’re a publicly traded company, and there’s nothing magic about them in terms of mergers and acquisitions.
As you consider all of this, how do you think it will play out over the next five or ten years?
Bernard: There are a lot of interesting dynamics there. One thing to remember is that when Citrix bought XenSource, they actually bought a proprietary company. Xen is open source, and there’s a kind of wrapper around it that was a licensed product. But the question, more broadly, is what it means when a proprietary company buys an open source company.
I sometimes wonder if these proprietary companies might one day wake up and unwrap the present and go, “Oh my God. We don’t like what’s in this box.” Because the thing about open source is it’s kind of a one way decision.
You can’t say, “This was open source, but now we’re going to go back.” If you own the intellectual property, you can choose to license it differently, but generally speaking, for example, if Oracle had bought MySQL, they’d have an open source database.
And the user perspective is quite interesting as well, because one of the things that open source provides for users is a hole card that says, “If somebody doesn’t do what they ought to do with this project, I have an option. I can take the code and go build a community around it on my own, or I can find somebody else who will do it.”
And truthfully, let’s just say Sun completely screwed up MySQL, or tried taking it proprietary or whatever. Somebody could then say, “Well, let’s take that same code and go start our own fork on it and go do something with it.”
From the user perspective, you’ve got an option that, with proprietary software, you never had, because the stories are legion of, “I chose XYZ product and got bought by somebody, and now they’re canning it, or they’re ruining it, or they’re not supporting it anymore, or they’re not enhancing it, and I’m stuck.” At least with open source, from the user perspective, you’ve got an out. It may not be the out you’d like, it may not be the out that you think is most attractive, but at least you’ve got some more control. You don’t have as much lock in.
At a broader, macro level, what this means to me is that we’re going to see software margins come down, because ultimately it’s going to have the impact of lowering prices. The availability of a free thing has the impact of making overall prices average down.
Scott: Another thing that I find very interesting around all of that is that when you buy a proprietary company, you acquire all their intellectual property. When you acquire an open source company, though, what you’re really acquiring is the talent of the people who know that code base, but you’re not really acquiring intellectual property.
Like you said, the code’s out there, and anybody who wants to can fork it tomorrow, or the community can just drift over to something else.
On one hand, you would hope that maybe Citrix would acquire Xen, and the open source methodology would be infectious, and it would spread throughout Citrix and they would move more and more products that way. Sun acquiring MySQL made sense, because Sun is very actively trying to open source all their code, and so it’s very unlikely that they’d try to take it closed and relicense it, because they’re trying to move everything the other direction anyway.
But Citrix acquiring Xen is fascinating, because the founder of XenSource built a hypervisor, built a company around it, sold the company to Citrix, and then went off and said, “You know, now that I’ve created a hypervisor once, there’s this other way that I think has merit, where I do it right inside of the Linux kernel, and he founded Qumranet and wrote KVM.
And so, what Citrix bought may not have a lot of value, because the open source community just kind of drifts to something else. There’s this very kind of Darwinian thing that happens with technology, where the open source community seems very willing to say, “Well, that was a good technology for a while, but this other thing really seems to have some advantages, and we’re going to start moving in that direction.”
And even with MySQL, Sun acquired it, but now Brian Aker is a Sun employee working on a fork of MySQL called Drizzle, which is sort of going back to MySQL 4.1 and ripping out a lot of the stuff they put in for 5.0 and re architecting it. I wonder if it’s a cautionary tale for companies that fork over a billion or half a billion, thinking they’re “buying” an open source project, only to realize that these are a lot more ephemeral things than when you buy a closed source company and acquire all of its IP.
Bernard: Right. Citrix bought XenSource. They didn’t buy Xen. And XenSource wasn’t in a position that MySQL was where MySQL really held the copyright to all of that stuff. XenSource was a company built on top of an open source project, for which they were very significant contributors, and maybe even sort of set overall guidelines, but they didn’t own that open source stuff.
I mean, I think what you’re getting at–and this is kind of a different way of saying lock in–is that people who use open source have very high standards and expect to be satisfied. And if you don’t continue to meet that hurdle of what they consider to be their satisfaction rate, they’ll go to something different. They can go off to a different project that doesn’t have that same characteristic of control over them.
It’s kind of a cliché within the open source companies that you have to be really, really good every day, during every interaction the customer has, because if they don’t like interacting with you, they can stop doing business with you.
It’s kind of the flip side of lock in, and in that respect, it is ephemeral. I’m not picking on Citrix, but let’s just say they started really dissing Xen and saying, “Well, XenSource is great, but Xen sucks,” or “XenSource is great, but Red Hat isn’t very good,” or something like that. There’s no constituency that has to stick with them. And so, in a way, it is ephemeral. It’s kind of a very different sort of dance from the traditions of proprietary software, for sure.
Scott: I think the other poster child of that is Netscape and Firefox. Netscape kind of open sourced what they did, but they weren’t really good stewards of it. Like you said, you have to be great every single day. And I think they weren’t.
And what happened was some people took the code base, they whittled it down, and the Firefox web browser was born. And you’ve got the Mozilla Foundation around it now, and that happened because, basically, Netscape wasn’t a great steward of that code, and somebody with enough gumption and enough technical ability just took it and ran with it in a different direction.
I think that’s kind of terrifying, from the perspective of a company, to think of that as being a possibility. But from the perspective of users, and people who are going to invest time in it, I think that’s a very reassuring thing, that if a company manages it well and they’re a good steward of it and they put a lot of developers on it and they keep moving it forward, that’s great; we all like that. If they don’t, we’re not necessarily screwed. Somebody smart might take it and go a direction that works better for more of us.
Sometimes even closed source companies can have great technology, and the technology just dies because the fundamentals of the company aren’t good: they’re not profitable, or there’s a lot of politics internally, or whatever. Open source provides a mechanism for great technology to live on in situations like that, where it can get re-purposed and turned into other things, regardless of what happens to the business entities wrapped around it.
Bernard: Open source represents an enormous transfer of control and power from vendors to users, and the implications of that are profound. What you’ve just outlined is a part of that.
There are many, many end-user companies that are in very dissatisfactory situations because they bought a product that may be good, but the company they bought it from isn’t. Or maybe the product is good, but due to the vagaries of something going on internally at the vendor, their product ends up being orphaned. And it’s a very frustrating situation, because then you’re out of control. And open source changes that power relationship.
Scott: I just want to say thanks for taking some time to chat with us. Do you have any closing thoughts on your end you’d like to add?
Bernard: Well, it’s been a great experience talking to you. It’s going to be a fascinating time, over the next five to 10 years, to see how the software and the IT industries evolve as a result of these factors, both from a user and vendor perspective.
Scott: Thank you.
