But Mark Spencer's company, Digium, has something that eluded many a Silicon Valley wannabee during the bubble: real revenues. The company pulls in about $10 million a year, and its open-source software has garnered interest from such business stalwarts as AT&T.
Spencer's company makes Asterisk, an open-source application, for the Linux operating system, that's at the heart of many installations of sophisticated corporate phone equipment. The upshot: Gear that typically costs hundreds of thousands of dollars is now available for the price of a laptop.
Digium generates revenues from licensing its open-source software to commercial interests. But it also faces new competition, a topic Spencer addressed during a recent CNET News.com interview, a talk in which he also discussed the impact of open source on telecommunications.
Q: In the past, you've described Asterisk this way: "The genie is out of the bottle and nothing will change that." What do you mean?
You've said your mission is "extreme capitalism." Explain.
Spencer: It's like a big game of Monopoly, but you can't let anybody win. Capitalism, to me, is a system of competition in which the end product becomes better and (less expensive) because of competitive forces pushing on it. Open source forces you to have to be more competitive. As you add competitive advantage, that advantage is adopted more rapidly; you can't sit back on your development. It forces you to constantly be innovative because people can use your new thing immediately.
More argument for a proprietary software--to monopolize your developments?
Spencer: But open source speeds up competitive cycles. You have to work very hard. With proprietary systems, sure, there's only one company that can implement any new development. But that's not scalable. With open source, resellers can make those changes on the fly and add those features.
Is the telephone industry ready for open source?
Spencer: There will always be room for proprietary software. But there are certain characteristics that predispose a market to open source. One is the size of the market; telecom is enormous. Secondly, there's got to be an audience with a high degree of technical skills. Telecom people are much more technical than most others.
What's the ultimate place for Linux in telecommunications?
Spencer: Look at what happened to the personal computer. In the 1980s, there was only so much you could do on a PC. Spreadsheets, word processing--it was fairly limited....Now supercomputers are running on AMD 64 architecture and Linux. Most of the proprietary architectures got pushed farther out to the fringe of technologies. That's what we'll see with Asterisk.
There don't seem to be a lot of developers writing PBX applications. So how can Asterisk grow new features very quickly?
Spencer: The technical nature of the audience makes the difference. In 1998, I created an instant-message client that's very popular, maybe 2 million users. But just a few dozen developers are contributing their work back into the open-source community. At 200,000 installations, Asterisk has 300 contributing developers.
You've authored your own Internet phone protocol, IAX, to compete with the very popular Session Initiation Protocol. Why?
But is anybody using IAX? And how so?
Spencer: Global Crossing uses IAX for its last-mile connection, and SIP for inside the network.
Two standards in one network. Isn't that proof of a standards schism?
Spencer: Asterisk's goal is to try and support a lot of different protocols. Different ones are needed to solve different problems. IAX makes a real good model for the last mile.
There are already companies saying they are making an Asterisk that's better than the original. So, in effect, doesn't your commitment to open source threaten Digium's very existence?
Spencer: What you're getting at is a serious issue. When we talk about open source bringing vendor independence and empowering customers to be in control of their own industry, it does mean we have our work cut out for us. The industry will be much larger than we are, and there will probably be companies selling more significant products. We can't sit back. If we didn't play our cards right, then the Asterisk wave could go on and we'd not be part of it as much as we'd like to.
How have you managed to generate revenues with open-source software?
Spencer: The debate about profits and open source versus proprietary is a moot one. You already can build a profitable company using open source. We've shown that. Now that it's there, you have to adapt your business model to the reality of the situation. Nowadays with the Internet, the idea of selling things on a per-license basis, as software's typically sold, just isn't compatible any more.
Any changes to Asterisk in the works?
Spencer: We have a business edition, which should be shipping this month. It's the same software put through more formalized testing, sold with a more traditional software license model and is for people not familiar or comfortable with open source.
Are you part of the new versions of Asterisk for proprietary operating systems, such as Microsoft and Apple Computer?
Spencer: We don't do that. But part of the reason we have a good relationship with our community is we play by open-source rules. So, when we develop new software, we do make it available. I don't anticipate that the versions for proprietary operating systems represent a threat. We want to be able to adapt to support commercial and open-source markets.