The Impact of Open Source VoIP

Late in 2000 my company launched one of the first free VoIP services in Italy, Chocophone, which was based on a proprietary protocol in order to improve the quality of phone calls over dial-up internet. In addition to this, the proprietary protocol let us make a very lightweight softphone without the need to carry all the stuff around the H323 protocol, definitely too “heavy” to work well over a dialup connection.

This choice led us to develop a server application where we had to convert our proprietary protocol into H323 in order to get connected to the most common and available media gateways to break into the PSTN.

What about the “white page” syndrome? Developing a H323 stack from scratch would have taken too much time. For this reason, we made a very deep usage of the OpenH323 libraries, something that saved us a huge amount of development time.

All that said, many companies contributed to the development of those libraries and made a deep usage of such stack. Developers had to be hands-on and use that code, make corrections, add features and so on and it was also a great way to learn about VoIP if you weren’t expert enough. At the end of the day, open source applied to VoIP already meant faster development, competence sharing, learning on the job, faster time to market.

Asterisk and SER have been and are, on the contrary, the icons of Open Source VoIP in regards of the SIP protocol. Many important carriers and ISP around the world offer */SER – based services to their users, leveraging the huge amount of development going on behind the scenes and the hard work and investments made by many players in the VoIP market in order to support those projects. In a nutshell, less entrance barriers to those innovative minds who want to setup their own VoIP company very quickly and saving money.

There are some downsides, of course, since when you start using platforms like those you are bounded to them with not that much space to integrate your own proprietary bricks to perform some very specific functionalities. On the other hand, you can do that by integrating your pieces of proprietary code to those Open Source projects, being pretty sure that you are working with widely used and standard implementations of VoIP and having the ability add value without the need to make your products open source too. Of course, a deep attention to the ongoing developments, branches, bugfixing and so on were/are a must.

The second part of the story of the impact of Open Source VoIP to the whole telecom market is made of softphones and stacks. Ekiga (formerly known as GnomeMeeting), Linphone or PhoneGaim, just to name some, have been a great source of knowledge for new entrant developers in this field and a great way for providers to offer a cheap way to make phone calls from your PC. SJPhone deserves a special mention too, even if it has been shipped for FREE but it’s not open.

Speaking about stacks instead, that is a complete set of libraries which include everything related to make/receive a phone call using H323 or SIP, those are a lower level piece of software which helped you to create brand new softphones without having to turn the main guidelines coming from RFCs into code. Vovida.org was a good source for developers in the past, even if it appears to be shut down now.

The next step is about APIs. Dameon aka Phoneboy wrote a very thoughtful piece for GigaOM which led to an interesting series of comments and other blog posts. In particular, I would like to highlight Dan York’s comment:

APIs need to be based, as much as possible, on *open standards*. Otherwise: 1) an API becomes just yet-another-way for a vendor to lock you in to their specific platform; and 2) developers can’t use their existing knowledge and have to learn how to use *your* new API.

I completely agree with Dan although I don’t think #1 VoIP vendors will ever embrace this approach because at the end of the day everyone offers its own implementation, definitely a good way to keep best developers and consultancy companies tightly linked to their brands and solutions. Right now it’s important not to reinvent the wheel and the focus must be on the application layer and in this perspective APIs are critical. Ask Ribbit or Tom

What’s next? First of all I would like to point out this comprehensive source of Open Source projects/resources. This area is in continuous development so this collection of projects might not be complete at this time, even if it is a great starting point. Secondly, I would like to invite you to join me to the Internet Telephony Conference and Expo in Los Angeles from Sept. 16th to 18th where I’ll be moderating a panel about Advantages of Open Source VOIP, with a specific attention to the Enterprise and SMB market. Trixbox, Voiceroute (Druid UCS) and Intuitive Voice will join the panel, offering their experiences and views thanks to the products they are proposing to the market and to the market’s response. I firmly believe it’s gonna be interesting

.

In conclusion, my last comment is for Skype. The main subject here is open source VoIP and how it has been critical to make the VoIP market as it is today. On the other hand, proprietary services like Skype have been fundamental to tell the masses what VoIP is and to make adoption explode. What does it mean? When it comes to consumers, they don’t really care of what this or that service use in terms of protocols, codecs, or whether it is open source or not. They care of usability, costs and it must work. Period. Not SIP compliant, far from being open source, but Skype is the best example of a winning service for the consumer market. The question is: what about businesses? I really would like someone at Skype to join us for the mentioned panel… I should drop TMCNet guys a line…

Comments, experiences, tips to share? Just leave your comments and I’ll be glad to join the discussion.

Tagged as: dan york, druid, ekiga, gigaom, intuitive voice, linphone, luca filigheddu, open source voip, openh323, openser, phoneboy, phonegaim, ser, sjlabs, sjphone, skype, tmcnet, trixbox, Voice 2.0, voiceroute, VoIP