Mark Spencer was just 21 when he first created Asterisk, an open source PBX (private branch exchange) that runs...
on Linux. He eventually realized that Asterisk could serve as a replacement for telephone switchboards, and that it could support VoIP (voice over IP) applications from phone calls videophone. Today, Spencer, the founder of Digium, boasts over 200,000 Asterisk users, many of whom contribute new functionality on a regular basis.
SearchEnterpriseLinux.com recently got on the phone with Spencer to find out more about Digium, Asterisk and the growing market for open source voice and data applications. Here are some excerpts from that conversation:
How far along are open source voice and data applications in terms of acceptance in the marketplace? What types of companies are using those applications?
Mark Spencer: It's catching on. It tends to be in different places. I would say that a lot of big companies see Asterisk as a way to build a particular niche. They've got some application that their existing infrastructure doesn't really fill, and they want to use Asterisk to fill that. A lot of smaller and newer companies tend to look at Asterisk as a very inexpensive way to create their infrastructure.
Are we talking only about VoIP infrastructure here?
Spencer: We're talking about the regular voice infrastructure. Whether or not you go with VoIP is entirely secondary to Asterisk because Asterisk supports both traditional TDM (time division multiplex) as well as the regular VoIP. You can deploy Asterisk in a traditional TDM application today and be ready for VoIP immediately. Or you can go directly and take that plunge into VoIP right away. Asterisk is kind of the bridge between those two.
What do you say to people who think that the telecom market isn't ready for open source?
Spencer: I see open source and telecom as a good match for a several reasons: There's, obviously, a huge cost difference between open source and proprietary solutions, so that's one factor. There's a very large market for telecom, so there's a lot of people that can use it. That helps. And the audience for telecom tends to be a very technical audience, which means that as a percentage of people that use it, the number that actually contribute back code and make improvements is very high. Of course, by 'high,' we're still talking about only a fraction of a percent. On Asterisk we have over 250 contributors, and probably about half a dozen that really periodically are putting in patches daily or weekly.
Are there any other reasons why you see the two as a good match?
Spencer: One of the big reasons also is that telecom is an area where there is just a lot of demand for customization. People want things to behave exactly the way they want them to behave. And that kind of a need is obviously compatible with the way that open source works in general.
For example, we have an application for Asterisk which is an alarm receiver. It receives and controls signals like you have in an alarm panel in your house. That would never come up in your roadmap. It's not something where I'd say, 'Gosh, I bet there is a really huge demand for that.' It's something that we were able to get because someone was able to contribute it. It fit their need. But we would never have been able to pursue it ourselves.
How do you go about supporting those firms that don't have open-source skills in-house?
Spencer: We've built up a reseller and distributor channel that is able to help people that don't have that resource. The resource of knowing how to do Linux configuration is not a really uncommon one. And what we've found is that Asterisk has enabled a lot of VARs (value-added resellers) to be able to get into the voice market where there can be a lot [higher profit] margins for the reseller. That's because there is a lot of money in voice right now. You've had this transformation of the data business with Ethernet. You now longer have to do everything with big and expensive proprietary stuff. That has been happening for quite awhile and has pushed down a lot of margins in that data space. People look at voice and say, 'Wow, when I go sell this customer not only can I sell them a Web server, but if I can sell them a PBX then I can make a lot of money.'
How does Digium make money?
Spencer: [One way we make money is that] we sell interface cards that connect with Asterisk. We also provide support and services. And we also sell licensing outside of the GPL (General Public License). So, if you need Asterisk in a proprietary application we can give you that capability.
We're also working on a version of Asterisk that we plan to introduce at the very beginning of next month called Business Edition. We've taken a snapshot of the code and run it through formal regression testing, fixing any bugs that we found along the way. We'll be selling that with more traditional software licensing and warrantees.