CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

Betting the farm on open source?

RealNetworks CEO Rob Glaser insists his streaming-media company will not suffer the same fate as Netscape as he tries a radically different tack in a bitter competition with Microsoft.

If the parallels between RealNetworks and Netscape Communications were once suggestive, they became flagrant Tuesday as the struggling streaming-media company followed the defanged browser pioneer in releasing its source code to the world.

Fewer things get deeper under the skin of Rob Glaser, founder and chief executive of Seattle-based RealNetworks, than the Netscape comparison. But both companies wound up losing market share to Microsoft in a crucial Internet medium after the software giant bundled its competing program with Windows. What's more, the RealPlayer creator and Netscape each subsequently pursued content-oriented strategies as well as open source plans.

Still, Glaser insists the similarities are only skin-deep.

Unlike Netscape's Mozilla project, or other less-than-successful open-source releases of formerly proprietary software titles, RealNetworks is going the open-source route from a position of strength, according to Glaser. The company is also responding to overwhelming demand for its products by partners who want RealNetworks' technology on their cell phones and IP (Internet protocol) devices, Glaser said.

So far, open-source advocates have given RealNetworks' plans high marks. And for the most part, analysts have concurred with Glaser's interpretation that the corporate open-source failures of the past have little relevance to his company's release. But questions remain: How long can RealNetworks continue selling a streaming server while Microsoft gives one away for free? To what extent can the company rely on the market for IP devices? And even if RealNetworks is in a fundamentally different position than its superficially similar predecessors, has the company adequately studied those failures and figured out how to avoid its own?

CNET News.com spoke with Glaser as he prepared to launch his company's open-source release.

Q: What gave you the idea to release this code?
A: We had been talking for a couple of years with some of our partners because we always ran into this issue of having a partner that wanted to build to and add on to our platform...Nokia came to us at the beginning of the year, and they gave us the road map of the devices they wanted to add Real to. And it was 12 different devices, and in some cases they had different operating systems. You could wind up sucking up a whole development organization in serving one partner. Nokia's a pretty important partner, with 100 million phones a year, but we thought, 'Wow, we have to change the model.' So with Nokia we created a very flexible paradigm of how we would share source code and maintain compatibility.

If you are a clone of something where there's already an adequate commercial instance out there, then open source is not a panacea.
Is that when you realized this was going to be important?
That was a watershed within our organization. It got us to see that this isn't something that just works in the abstract, and when we got ready to bring the Helix products to market, where there were dozens of manufacturers with hundreds of devices, and that people wanted to have the Helix architecture included, this was the only scalable way to do it. And we put a lot of thought into it and wound up with what will be a very successful approach.

Was the idea controversial within the company?
Yes. I would say that we've had a few people in the company who were strong believers in open or community source. It's not as simple as to say that there were engineers on one side and business guys on other, because people on both sides have been tracking and following the open-source movement for quite some time, and have seen it move over four or five years from the fringes to something indisputably commercially viable.

RealNetworks isn't the first company to take its proprietary software and put it into open-source development. What did you learn in studying your predecessors, both in terms of what to imitate and what to avoid?
We learned from all the predecessors. From Sun we learned that it's possible to create a very vibrant community with built-in compatibility rules. With Linux--they're more organic in terms of the core group that's responsible for what changes get checked in and checked out. But we also learned that there are people who look at the Java community approach and say, 'I want to use this as a launching point for my work and am happy to put this back in, but I want more flexibility in terms of compatibility.'

You've referred to your open-source predecessors as successes, but many of the open-source projects that began as commercial, proprietary products have not thrived in an open-source environment, with Netscape being the prime example but also Apple with its QuickTime server and Sun with the StarOffice project. Why will RealNetworks' giveaway be any different?
I think that if you are a clone of something where there's already an adequate commercial (example) out there, then open source is not a panacea. By the time Netscape put Mozilla into open source, IE was already functionally equivalent to Netscape...If there's something that's already a commercial clone and you're not offering unique functionality, open source won't completely turn the tides. It's not like StarOffice had 40 percent share and Microsoft had 40 percent share. StarOffice had market share that rounded to zero when they put their source code out.

In our case, because the Helix DNA client is the only client that supports the Real format, this is the most comprehensive solution from a functional standpoint in terms of a media delivery platform. It's not like there's some other implementation out there that supports Real formats and the standard bodies' pieces. Helix DNA provides a unique value-add. This goes for the PC side, but also for the new category of IP devices.

It's been 13 weeks since you announced you were going to release this code. What has happened to that code in the meantime?
We created a more modular architecture than we previously had, and we did that by forking the tree of the source code to accommodate limited-function devices. With Helix you have modular profiles that let the implementer specify the footprint and the functionality they want. It's a more scalable way of architecting a system.

Your third-quarter earnings showed solid growth on the consumer side of the business. Where's the bulk of that money coming from? How will you maintain that growth, and what new content do you have coming down the pike?
We announced that the subscriber revenue grew 20 percent sequentially and that consumer growth was 11 percent. A majority of the growth came from the subscription piece growing. We are growing RealOne globally, with seven international versions--French, German, Italian, Spanish, Japanese, Korean and Chinese--which we haven't announced yet, but it's imminent.

In terms of partnerships, we just signed a college sports partnership with OCSN (Official College Sports Network), with 50 college sports. And those pieces are working, and we're seeing increased usage of the product and continuing word of mouth in addition to new content.

We're not focusing on the Scotch tape business. It's a much broader strategy.
With the Helix code release you've really emphasized the mobile-phone manufacturers and consumer-electronics devices. To what extent has RealNetworks been adopted in these environments, and what's your goal?
I think there are three fundamental trends that our company is optimized around. The proliferation of IP devices...the growth of broadband...the availability of audio and video content in digital form.

If you look at the history of new consumer-electronics devices, there are some big successes and also a high mortality rate. We wanted to support a very broad range of development because we don't know what's going to survive in the marketplace, and only through the open-source model can we, or any other company frankly, cover the waterfront.

In your third-quarter earnings announcement, you said, "On the systems side, our business remained challenging." Now you're announcing a shift toward a sort of telecommunications infrastructure gateway. What exactly does that mean and how does it help your systems side grow more rapidly? And how does it relate to the Helix release?
It's fair to say that in a model where we're making our core plumbing available to the community as a whole, and we have license fees that could accumulate and be significant, we're now going to have a server that is, by nature of being a community product, not going to be priced on a stream-count model.

So you'll see us move toward a solutions orientation in our systems software efforts. The Universal Server, as opposed to the Helix DNA Server, has a bunch of applications for network management, content management features. Anyone could build on top of the DNA server. We might wind up competing with our own licensees in that area, but our focus will be on building these applications principally for media and telecommunications industries, for wireless and broadband content-delivery networks, radio, television and film. These are huge industries. We're not focusing on the Scotch tape business. It's a much broader strategy.

You say you're planning to abandon the stream-count pricing model, but meanwhile, Microsoft offers its streaming server free with the purchase of its operating system. Is there any chance you'll follow suit and make the server free altogether?
We're going to continue to sell servers, and we'll still sell professional servers. The only question is which Helix product you buy. We've just announced the pricing for the DNA client. We haven't released the encoding or the Helix DNA server piece, so we will announce that toward the end of the year.

Do you expect to spend the rest of your life doing streaming media? As a rule, do you think it's a good thing for a technology company like RealNetworks to operate under the same CEO for five years? Ten? Twenty?
We started this company when I was the ripe old age of 32, and I just hit 40 earlier this year...While I revisit from time to time, every three or four years, my enthusiasm for what we're doing, and my interest in this kind of a role versus a more philanthropic position--and the board has the right to revisit it as much as they want to--I can't think of anything more fun and rewarding to be doing.