CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

Sun's open-source odyssey

Sun's incoming executive vice president of software says the company envisions a world dominated by open source.

Among the reasons Sun Microsystems has been out of favor with Wall Street for years is the company's inability to fully profit from its clever inventions, notably those in software.

With the company now pursuing another turnaround, the nub of Sun's software game plan can be boiled down to two words: open source.

The company has open sourced its Solaris Unix operating system, and CEO Jonathan Schwartz--a software man himself--said Sun intends to eventually open-source its entire software product line, even the Java programming package, something outsiders have long urged Sun to do.

The idea is to steal share from entrenched software providers by using the disruptive potential of open-source business models, where software is available for free and vendors charge for support services.

Rich Green, executive vice president of software at Sun, is tasked with making that plan a profitable reality. In May, he returned to the company after two years at a start-up and said he found a company that has shed its "religious biases."

Speaking with CNET News.com, Green said Sun will open-source Java "pretty quickly," and he described how the company aims to compete under the Darwinian rules of the software industry.

Q: Your primary mission right now is to speed Sun's transition to open source. Where are you in that process?
Green: When Jonathan (Schwartz) called and offered the opportunity to return, it just seemed to me that the company that I could return to was in a much different place than the company I left two years prior. It is now being led by an individual (Schwartz) who has a strong software bias, there were many advancements in terms of the open-sourcing of Solaris, the success of the NetBeans developer community, the advancement of Sparc architecture to the CMT (chip multithreading) technology, the realization that the world is not Sparc-only, and shipping X64 machines.
People don't buy things until they have obtained it, analyzed it, tried it out, manipulated it and maybe built software on top of it.

It just seemed to me like a company that had dramatically reduced its religious biases and been focused more as a business--and, in fact, a progressive business--on the current hardware-software-services lifestyle.

The knock on Sun for many years is that it is a hardware company--all it wants to do is sell more servers. Are you telling me that Sun has gotten a new religion and software is the lead?
Green: Well, these are shades of gray. I think historically our dramatic skew was towards hardware-only, and now I think it's a much greater balance. You know, we're a big company and we can afford to do numerous things well. So I'm not going to sit here and say, "We're going to software instead of hardware." But what I am going to say is that we're going to run software as a significant peer component of Sun's business, certainly use it to positively affect our system sales.

(We are) an entity whose charter is to go forward and compete aggressively in the industry with, I think, a pretty rich and capable stack of software at the run-time level and a set of powerful development tools. It's a much different perspective than it used to be.

You've said that Solaris is the centerpiece of the software strategy. Why is that? I think a lot of people might argue that Java has the stronger attachment, or brand association, with Sun as Solaris.
Green: I actually think you're right. It's hard; it would be hard to choose which is more important than the other. I will note though that in today's world, there is a lot of competition at the operating system platform level alone and so it's not as if Java is more or less relevant, it's orthogonal.

Everybody has always recognized the technical prowess or capability of Solaris. I think the accessibility...in the open Solaris program has allowed people to more viscerally appreciate it because they can see the code, they can use it, they can add to it, it's a different world.

Why does providing code lead to more sales?
Green: It's not only having the code or not...People don't buy things until they have obtained it, analyzed it, tried it out, manipulated it and maybe built software on top of it. So there is a shift (in software acquisition) from pre-analysis to post-analysis. It brings three audiences (developers, systems administrators and CIOs) into the mainstream of the analysis cycle because they go get it.

Whether they are reading the sources or just trying out some projects, they can do it at their pace with no intervention or interaction from Sun required. Then when it's time to go big, they give us a call. It's a much different model for Sun and for any other entities engaged in open source than the prior model of monetization at acquisition.

One of the first big announcements from Sun after you joined was that the company intends to open-source Java. Summarize what you think you'll achieve by doing that. Also, could you address compatibility, which we've been told has been the thing that has historically held Sun back?
Green: I want to separate out the buzz in the industry from the reality, which is, by and large, most individuals have full open-source access to the technology. That said, the means of licensing, the flexibility of using that open source is affected by what Jonathan and I announced at JavaOne. And that's why we're going to take the steps to go fully free up the technology. But it should be well-known that virtually everybody gets the access to the source (code) of Java today, but we want to further make flexible what you can do with that stuff going forward.

So, we fully intend to do it. It makes perfect sense. It kind of removes from the system the noise or the angst of Java in terms of access and flexibility. So that's a big deal. Now the compatibility issue is a risk, but I think it's a risk well worth considering taking. Not only is Java more advanced than, I think, any other open-source software in terms of compatibility testing, the availability of (testing suites) and other things like that, but the number of applications out there is so enormous that they tend to drive compatibility. I think this is a manageable issue.

You have to make sure that you don't get too hung up on history being the only way to do things.

When do you expect to actually open-source Java?
We're working pretty hard on making this happen pretty quickly.

How much time do you think you have before Sun shows results from its investments in software?
Green: You know, the results are multivariate. We did 5 million (downloads) with Solaris. So how're we doing? Not so bad for a year's work. The metrics are interesting: They are downloads, adopters, developers, financial goals, etc. I think there's a lot of acceleration in the system already.

I think where we really have to spend some time as a company, as an organization in the community, is in the middleware area in particular. Between the acquisitions we did for identity and business integration as well as the rest of the industry-standard open-source middleware stack, we have some really good stuff out there, and it's packaged in such a way that it is more usable than individual piece parts. I don't think we've done a good enough job of getting the message out, that that technology should be considered.

The announcement we made at JavaOne of open-sourcing that whole stack, integrating it with our developer program, getting more sort of eyeballs onto that technology is a big priority for me.

It seems as if all large vendors are appealing to developers. Do you think there is more competition for their eyeballs?
Green: In the context of open source, there are more lines of code, more artifacts for people to look at than ever before, I agree with that point.

You know, you can argue (that) we are not beginning to do this, we are returning to our roots of it, and in that regard I think there's a fundamental deep respect for Sun with developers and administrators with regard to our core technology, our technology strengths and our open-source plans. I think once people got through the "are they really going to do it?" (question), I think we're going to end up very quickly on the No. 1, No. 2 list of companies you look (to) for open-source business models and technologies in the industry; it's 20-plus years of us doing this.

You see a lot of activity among developers happening in open-source projects and outside the standards processes, where most of Java development has historically happened. Is this a good thing? Is this a bad thing for Sun and Java?
Green: Oh, I think it's a great thing. You have to make sure that you don't get too hung up on history being the only way to do things. Standards were a great way of operating in the industry in a pre-open-source-world lifestyle, because they were the only way to gain sort of visibility and normalization or compatibility in products that were available in binary form. Now that things are available in source code form, (there are) different models of innovation and creativity and different notions of what is standard.

So we're not trying to control it. We're not trying to say, "If it's not Java, it's not good." You'll see us reaching out to these projects and programs and supporting those things in ways greater than we've done before.

Scripting languages is one of the hot areas in development right now. What can you do to appeal to that crowd? Some people argue that the Java virtual machine should be the .
Green: I actually don't want to see things go in any way that the industry doesn't want to go. So I'm not trying to dictate a direction, you know, I think the fact that Sun hired Tim Bray...the most-read Web 2.0 scripting guy on the planet, is an indication of our intent. He is the guy who was so in touch with this community, or rather these communities. He is a fan of Java, but he's not only a fan of Java, you know, I think in many respects running these (scripting) environments on top of a virtual machine is a wise idea, but Darwinism reigns. We'll see what developers do, and you'll see more and more programs and energy offered up to help them innovate and help them decide. We're not going to lead the witness here.