Martin Fink's book on the business side of Linux and open source was one of the first books that looked at Linux and open source from the perspective of corporate managers and executives. In this extremely influential book, he explained in business terms why Linux and open source are here to stay and why companies everywhere should be aggressively moving to develop strategies for their adoption. He now speaks at conferences around the world delivering this same message.

Martin has emerged as one of Linux's leading voices in the executive boardroom. We at LinuxWorld Magazine are excited to be able to bring you this exclusive interview.

Thanks so much for being with us today, Martin.
My pleasure.

Your book, The Business and Economics of Linux and Open Source, has been widely praised by managers at companies around the world. How have you seen the reception of the book?
I've actually been very pleased with that because the whole reason for doing the book was to document what I think has really not been documented before, which was targeting a management audience so they could understand not just the technology, but also how it applies to their business and whether or not it would fit in their environment to allow them to either increase productivity or lower their costs.

For our readers who haven't read the book, what are some of the major themes that you touched on?
First, for those who are completely new to the Linux and open source environment I start off just doing some basic background education because there are really some new paradigms and new concepts to think about with Linux and open source in understanding licensing methodologies, etc.

The second major part of the book looks at how to deploy Linux and open source software within a company's enterprise and how to really do the cost analysis to say "does this make sense for me?"

The last major section is really targeted more at those who are software developers or large IT enterprises who do much of their own internal development so that they understand the dynamics of the organizational structure of the open source movement and how they can integrate either their company or their IT into that community.

It's been a couple of years now since the book came out. What have you been doing lately?
You're right. It has been about two years now since the book's come out and there have been some changes and some consolidation in the industry. I think the IT industry in general is much more comfortable and educated around the Linux and open source movement and understand it very well. I think that, just like any industry, we move from one phase to the next.

I think the first phase was people understanding what this new paradigm was about, and now we're moving into another phase of how to implement things and how to deal with some of the latest challenges. Within the open source movement many of the challenges we have to spend time on are in the area of intellectual property management, so I've been spending time making sure that those are well understood and people are making informed choices.

What are some of the challenges in intellectual property management that you're seeing?
One of the challenges is open source license proliferation. Any piece of software, whether proprietary or open source, is actually licensed for use to the end user and many people still believe that open source really is one particular license.

What has happened over time is, because people have taken different approaches to trying to license their intellectual property, a vast number of licenses have been created - I think we're approaching 55 licenses now. In my view, that has started to cause confusion as to which one of these licenses are appropriate to deploy within an enterprise. I'm looking to try to drive a simplification of licensing to make it easier for people to deploy and make it through their own legal department.

Another interesting area is digital rights management. Many people think about digital rights management as just protecting movies and music, but it's just as much about protecting corporate documents, corporate e-mails, and those kinds of things. I'm really trying to focus a lot of energy in making sure the technologies that are developed to protect our digital content are compatible with open source and open source licensing, so we can have a community that's available to develop and companies are not locked into proprietary technologies. Those are two good examples of where I try to put some energy around IP management.

I read a press release about a year ago where Hewlett-Packard announced they had over $2.5 billion in Linux-related revenue. How does that work since you don't actually sell Linux directly?
What we did was an internal analysis of Linux-related revenue. For example, we sell a significant number of server products that run Linux when they reach the final customer. We also have storage systems that run Linux. We deliver a significant amount of support and consulting services. We have a number of our printing technologies, or actual printers, that are connected to Linux systems. What we wanted to do was to be able to look at the very broad and extensive portfolio of products that HP has and, using a number of different methodologies depending on the product area, try to understand what is the Linux-related revenue in each of these areas.

It really sent a strong message that Linux is a very, very strategically important business for us.

How would you say working with Linux has changed HP as a company?
We have a term we use inside of HP called "customer-focused innovation," and I think that one of the key areas where Linux and open source has made a contribution is really getting us to ask ourselves, "Are we innovating and adding value in areas where customers perceive value?" and we define that as "Is a customer willing to pay for these capabilities?" We try to separate ourselves from things that might've been interesting, might've been cool to do, but really aren't perceived as adding value for customers. I think Linux has done a great job of highlighting that for a variety of different businesses in HP.

I know a lot of companies are looking to open source some of their own internally developed software. Is Hewlett-Packard looking at doing that and, if so, what types of criteria would you look for?
HP has been doing that for quite a few years. I am responsible for approving all of the open source contributions that we make or anything that we take in as open source software, because we actually do both. I approve now, on average, about three projects a week. These can be anything from releasing an entire product to open source or it can mean a simple set of contributions to an existing open source project.

A great way to learn about some of those contributions is to look at projects we have listed on a Web site called www.opensource.hp.com. There you'll see projects such as OpenSSI [the "OpenSSI (Single System Image) Clusters for Linux" project - ed.]. That's a good example of a large project where HP is making a contribution.

We ask ourselves questions like, "Can we make a contribution and create a community around this open source software? Will people come and participate? Are we trying to create the standard so that the industry gels around a particular standard?"

We have a number of criteria like this where we have the opportunity to ask, "Are we going to create a community and a standard and then be able to add value over and above the open source software?", so we can use that value to generate revenue for HP and for our partners and make our customers feel comfortable that they have the best technology.

A lot of companies are open sourcing different products. How do companies justify that when it's a product that could be generating revenue?
One of the things to be careful about is that open sourcing is not a panacea and it's not an answer to all problems that may exist in a particular product. One of the mistakes that some companies make is thinking that simply open sourcing will cause magic things to occur and all of a sudden they'll reach a dominant position because they are associated with this particular open source project.

HP has made its share of mistakes. We once released a project to the open source community and nobody came. That project was highly unsuccessful because at that time, which was five years ago now, we hadn't really matured enough to understand the dynamics of open source and what effect it would have.

Today we're much more disciplined in making sure that when we open source some software, a community will gel around the particular project. It doesn't have to be as large as, let's say, Linux or Apache. The value of open source is in the building of that community.

What do you think of the Open Solaris project where Sun is making their Solaris operating system open source? That's a pretty big product that generates a significant amount of revenue for them. Do you think that that's a good decision?
Time will tell, but my initial view going into this is that it will not generate the results that Sun is hoping to achieve. There are a couple reasons for that.

One of the base rules that I suggest people think about before they open source a project is: "Are you trying to compete against an existing open source project?" And really, in effect, what Open Solaris would be attempting to do is compete against Linux. There is already a vibrant community of thousands of developers who have gelled around Linux and there really is no compelling reason why someone would stop working on Linux to go do something with Open Solaris. I think there will be a significant challenge in creating a relevant community around Open Solaris.

I think a second problem is that we have not yet seen the licensing model that Sun is planning to use. They so far, at least in history - and we can only go based on what they have done so far - they really have struggled with truly adopting open source licensing methodologies so those challenges have been prevalent in things like Java, for example.

The third issue is that the version of Solaris that is being open sourced is not the Solaris that runs on their Sparc architecture; it's the Solaris that runs on their x86 architecture. There really are very few applications and very few customers who actually use that version of Solaris. When you reduce the number of users, then you've reduced the development community and you've reduced the ISVs that are there.

All of these things don't add up to a success story. And then you just have to ask yourself the question, "If I already have a Unix-like open source operating system [i.e., Linux], why do I need another one?"

If we take a look at these ideas now from the perspective of a Linux user, how does it change what they should expect from their software vendors?
If we really oversimplify this, I think that what it boils down to is more functionality at a lower cost. The more open source technologies that are developed over time and made available, the more technology that vendors can incorporate into their products. This allows the vendors to focus their energies on more and better functionality higher up in the value stack.

The really good news from an end-user's perspective is that open sourcing a given technology sends a pretty strong signal that that technology is in the process of becoming a commodity. As a result, they can expect the cost of those components to decrease over time.

Can you describe some of the contributions that Hewlett-Packard has made to the open source and Linux communities?
Our contributions can really be looked at in two dimensions. One is in actual code contribution, but another dimension is the investment we make in either the people or the organizations that are involved in open source. Let me try to intermix some of these to give some examples.

For example, we took our printer technology - HP is the number one printer producer - and we created a set of device drivers so those printers would be compatible with the Linux operating system. When we looked at the best way to deliver that to the community, open sourcing those drivers and letting the community enhance them really made the most sense. That's a pretty visible and very well-known contribution that we've made. This is just one example of many areas where we've contributed code to the open source community.

We've also invested and made a number of contributions to organizations such as the Free Software Foundation; the Free Standards Group, which is responsible for key standards such as the Linux Standards Base; the GNOME Foundation and the KDE League, which are the user interfaces responsible for the desktop area of Linux; and Linux International, which is an international consortium. We spend quite a bit on investment in the Open Source Development Lab as well; I actually sit on the board of directors of the Open Source Development Lab.

We also have a number of employees who are quite prominent in the open source community. Among them are people like David Mossberger, who is the core author of the Itanium Linux kernel, or Jeremy Allison, one of the key maintainers around the Samba project, which is a quite popular open source project. Another example would be my chief technology officer, Bdale Garbee, who has spent some time as the project lead for the Debian project.

I'll stop here, but that represents a fairly small cross section of all of the different areas of open source investments that HP has made over the years.

Is there anything else you'd like to share with our readers?
I think the most important thing to remember is that Linux and open source are not a cure-all. They're not the thing that's going to solve world hunger, as they say, but rather another capability that is available to an IT manager to increase productivity, lower cost, and get better functionality in their environment.

Thanks again for speaking with LinuxWorld Magazine, Martin Fink.

About Martin Fink
Martin Fink is vice president for Linux at Hewlett-Packard and author of The Business and Economics of Linux and Open Source (published 2003 by Prentice Hall PTR / ISBN 0-13-047677-3).