Asterisk is an open source private branch exchange (PBX) originally created by Mark Spencer of Digium. A PBX, is a type of phone switch, that allows multiple attached telephones to make calls to one another, and to connect to other telephone services including the public switched telephone network (PSTN). Not totally unlike routers that connect multiple computers to a network.
As the guest speaker at the Triangle Linux User’s Group (TriLUG) in Raleigh, NC on Thursday Mark recounted how he built the open source Asterisk project and the for-profit company Digium. He spoke about the journey from the open source Asterisk project to a company that provides support and services.
Mark got started with Linux early in 1994 Slackware (kernel version 1.09). One of the few in Auburn, AL at the time that knew anything about Linux. After a stint with Adtran (a global provider of communications equipment) he moved out on his own starting a company called Linux Support Services. Mark traded his services for space in the back of a local computer store in Huntsville, AL. He started by building his own computer and eventually built his own phone switch informed by his experience developing GAIM (now Pidgin) among other open source projects.
After explaining his need for capital to his friends at Adtran they offered to invest in his company. Eventually, he realized that he was receiving more interest in Asterisk PBX than for his general Linux consulting services. Then Mark met Jim Dixon who was building open source hardware. Their first project was to build an open source T1 card. This revenue was keeping them afloat but they didn’t receive any contributions and others were just taking their design and manufacturing cards that competed with theirs. Ironically, the same thing has happened with Asterisk though even with amble opportunity he chose not to talk poorly about those who some may say hijacked his work.
Money was tight at Digium until one day a salesman from DeltaCom (a southeastern competitive local exchange carrier) walked in to sell Mark and Jim a T1. After understanding what Mark and Jim had built the salesman offered to help them out gratis. From that point on they started seeing a steady increase in sales, and ended the year with a profit. After living on a meager income for so long Mark was able to grow the business without dipping too deep into the profits.
The folks at Digium soon realized that selling to technical users was good but they knew that they needed to package the product so that it could be easily used by a much broader less technical audience. As a result they improved the Asterisk web administrative interface and took steps to make sure that Asterisk was attractive to a much broader audience.
He attributes their early success contributions from this technical audience with contributions from hundreds of developers and feedback generated by over a million downloads in 2007. They also benefited from the fact that traditional telephone systems were very expensive and features didn’t translate from one system to another. Asterisk was very cost effective, configurable, and capable of VOIP in an industry with no clear leader. Also there was a fundamental shift in the industry from hardware switches to software switches.
The Easy-to-Use Asterisk Web Interface.
To connect a phone switch to the Publicly Switched Telephone Network (PSTN) required expensive training and licensing. The Asterisk project benefiting by Digium’s support to help overcome the the hurdles of testing and certifications. Digium also enabled their community by licensing G.729 a voice compression algorithm used for voice over IP on low bandwidth networks.The upfront licensing of this codec would have been cost prohibitive for most open source developers so Digium took the licensing hit and legally resells the codec for $10. All keys to open source success, enabling adoption in their open source community.
At the end of 2006 Digium was profitable had 80 employees and no budget Mark simple signed checks on the spot depending on company needs. There was not a single vice president. He was in a unique position with a multi-million dollar company and no pressure to take outside investment giving him a strong negotiating position if he wanted money.
Eventually he did take venture capital so that Digium could expand. Having a successful business with customers and profits allowed him to maintain control. He had a provision in his funding contracts that made sure he remained president of the company. That was short-lived. He realized he had little desire to be the president and in January 2007 he hired his old boss from Adtran to be the CEO of his company. He described his reasons were so he could still work with the technology and maintain the company’s culture.
When Mark started Asterisk he did a very smart thing. Every developer who contributed code was required to sign an contributors agreement so that copyright was assigned to Asterisk and pledge that there are no encumberances on the contributed code. This allowed him to be comfortable that his project was completely open source and that his company could relicense the code to OEM vendors like 3COM and NTT. Digium also has done the right things in keeping the community version fully featured and not creating a rift between themselves and their supporters.
One interesting note from Spencer on cell phones. He contends they were the best thing to happen to the emergence of voice-over-IP because it lowered the bar for voice quality and allowed VOIP to be tolerated as it entered the market. This is a great demonstration of disruptive technology because Asterisk wasn’t better than a hardware phone switch with copper they were just an interesting high value alternative. Despite the technical capabilities and improvements in VOIP very few people are building solely VOIP solutions today most Digium customers are using traditional landlines only partially doing VOIP.
Spencer also notes that Digium, a small company, is approached by large enterprises all the time. He notes that even though technologically they can provide a very reliable, robust product that some perspective clients don’t believe that. He even jokes that sometimes they have to tell prospects that some of their features don’t work very well to make them a little more credible as their true capabilities are somewhat unbelievable.
More than half of Digium’s business is done abroad. Even as the world starts moving to VOIP Spencer notes that TDM is far from dead. Many countries are still limited to much older technologies. When asked about the future of the Asterisk
Mark notes these would probably include a wider variety of messaging options including using Jabber or other technologies. One of the transitions he’s had to make mentally is that three years ago Asterisk was hot, now it’s becoming an engine that other products are being built upon. This is a real sign of open source maturity. Innovation will continue to flourish around Asterisk but probably far beyond the walls of the Huntsville, AL Digium offices.
His point on innovation is that improvements and innovative features for Asterisk are a result of problem solving by third parties. One humorous example is the The Booty Dialer. The dialer goes through your "bootycall list" and cycles through a list of numbers until one of the numbers gives an automated the reply they are available. However this is only one of a few examples.To promote future innovation Digium sponsors the Digium Innovation Award which recognizes developers, customers and partners for outstanding achievements that are improving business processes, overcoming technology challenges and enhancing the company’s bottom line.
User led innovation is one of the most powerful aspects of open source development. This includes things like Firefox extensions and SugarCRMs Sugar Exchange. Rewarding and marketing Asterisk innovations is a brilliant move on their part.
When asked about the future of open source hardware he isn’t convinced it will work in the same way as open source software. He cites the the barrier to entry for hardware production versus the rather low entry point to develop software. One example that he does see working is the the Open Cores Project that run on Field Programmable Gate Arrays(FPGA).
At thirty years old Spencer looks the part of the new tech titans like Kevin Rose (Digg) and Mark Zuckerberg (FaceBook). However, his company is making a real tangible product that addresses critical business needs. I suspect that Digium will continue to grow as they displace older technologies in telecom. I also think there’s real potential for other companies to drive their success by creating an ecosystem of add-ons that only increase Asterisk’s popularity. As someone who used to buy big telecom I am fascinated how effectively and inexpensively you can implement solutions that match the functionality of million dollar equipment for a few thousand.[Photo of Mark Spencer courtesy of Tanner Lovelace.]