Mark Spencer, CEO, Digium/Asterisk

What has been happening within the Asterisk space over the last 12 months?

Ans: Within the Asterisk space we have seen a lot of growth in areas like conferencing. In terms of development we have seen lot of improvements in conferencing engine. We have seen more people pushing Asterisk into larger kinds of installations with more centralized control. We are seeing these to take off in terms of doing peer-to-peer Asterisk networks. Those are the big trends that we saw over the last 12 months.

In general we are seeing Asterisk being used in lot of different kinds of installations across the board all the way from carriers to large enterprises to medium enterprises to small office PBX. Its map looks a lot like the map of Linux ranging from embedded devices upto super computers. It is very very broad. Another interesting that we saw in 2005 is more lightweight installations of Asterisk than we have seen before.

What will be the impact of open source resources, such as Asterisk, on the VoIP market?

Ans: The VoIP market in general is ripe for open source. There are also some characteristics of the telecom market that make it ripe for open source. First of course it is a very large market. Secondly the direct audience for products is a very technical audience. As users they are more likely to be able to contribute. Thirdly, telecom is an area where there is a lot of demand for customization. There are lot of business models to support open source in this environment. The other thing I think sometimes gets lost is that when we look at the telecom market, a product like Asterisk faces a lot fewer barriers to entry than for example a product like Linux does in computer space. People did not know about open source. You obviously had the monopoly of Microsoft, which had to be overcome. People were very familiar with windows interface and they were resistant to change and you had all these applications that only run on windows. There is nothing like that in the telecom space. There is no host of applications that only runs on one or another particular PBX or VoIP gateway. The standards are fairly easy to come across.

How does Asterisk compete with the IP PBX products made by reputed vendors?

Ans: The tight integration between the TDM side and VoIP side makes Asterisk unusually well suited to transitioning from TDM services to VoIP services. Also Asterisk has ability to run either in a very distributed fashion as is normally done in large networks where you have separate servers for voice mail and conferencing. It also can run in an integrated fashion. It is appropriate for embedded platforms.

Your Asterisk platform runs on Linux operating system. What in your opinion is the potential for Linux in VoIP?

Ans: Linux is obviously a natural compliment to Asterisk because it is very valuable to have an open source operating system. These two to a great degree go hand in hand. We have a lot of driver framework that we have developed. Linux is almost becoming a defacto standard in open source. Open source project scales as long as you keep good flow back of contribution and keep the development pace. I would like to call it “Extreme Capitalism” because with open source projects you can take code from one another which forces the pace of development to grow much much faster than in other situations.

With all the development and contributions towards Asterisk, what are the challenges that remain?

Ans: There are some feature related areas where a lot of us have interest and we are working on of course. We are working on adding video support. We did a technology demonstration with Grandstream using their videophone to be able to record and playback video. We have started working on integrating support for Google Talk protocol (Jingle). Google service runs on a different protocol. It does not run on SIP. It is built upon Jabber. When I first heard that Google was using another VoIP protocol, I was  questioning why they would create another VoIP protocol when there are already so many. Then when I looked at it, the protocol is very simple and very tightly integrated in with Jabber. It made lot of sense as to why they ended up going that route rather than the traditional SIP route. This is another area where Asterisk can benefit from its architecture because it is designed to support multiple protocols. We are working now to add support for the Jingle protocol.

Asterisk also has its own IAX (Inter Asterisk Exchange) protocol. How does you own IAX protocol compare with SIP?

Ans: IAX has certain advantages over SIP in certain areas. It depends upon what you are trying to do. The scope of what IAX does today is certainly far less than scope of what you can do with SIP and all the associated RFCs. IAX is good at just making a phone call. It does not suffer from NAT issues. It does not have the complexity in terms of call set up that SIP has. It has integrated dial plan support. The end point does not have to know about the dial plan. It has number of features that make it very easy to deploy voice services and easy to support voice services that make it valuable in number of different installations. We do not think it will be a SIP replacement as much as we see a world in which there is a need for more than one protocol because not every problem in voice is the same. If you look at email, there are SMTP, POP3 and IMAP. These are three different email protocols that are in use today. Each of these protocols is designed to meet a little bit different need even though all three have to do with moving email from one place to another. They follow three different models.

What is this DUNDi protocol about?

Ans: It is an open source peer-to-peer protocol. We expect DUNDi to be used in enterprise where you can use it on a smaller scale to be able to create separate PBXs that have no centralized point of failure. You can also add more resources by adding more machines to the PBX. The second is carrier space. There are a few different models for it. One is to try to push DUNDi all the way down from top to bottom. The other is to use DUNDi in parallel with technologies like ENUM and other related service.

Legacy vendors have not historically allowed resellers to customize solutions, making it harder for the resellers to differentiate themselves from their competitors? What role can open source platform play here?

Ans: I would say that the channel is a very critical part of why I think open source will be successful in telecom. Open source naturally provides resellers with a much more unique opportunity to differentiate their product. It is not just the same as pushing a box through. With customized solutions, the customer cannot just go out and buy exactly the same thing the reseller makes.

What does a platform like Asterisk mean for legacy PBX vendors, and how can they leverage it?

Ans: I think lot of vendors are still trying to decide what Asterisk means for the business, how they are going to use Asterisk or try to compete with it. So far we have not seen as many debates as we would have expected. When Linux was out there, companies were ignoring it for a longtime. Then Microsoft finally started trying to say bad things about Linux. We have not seen that for Asterisk from the proprietary telecom vendors. So far we have not seen bad mouthing coming from legacy vendor community. Obviously Asterisk can be both competitive and complementary to lot of products that are out there.

There are other open source telephony platforms such as OpenPBX and Pingtel (SIPfoundry). What are the differences in approaches here?

Ans: They are sort of different. If you look at the Pingtel offering, they took a proprietary PBX and then they tried to make it open source, which is a difficult route to go. If you compare the activity in Asterisk space with the activity in Pingtel space, it is very different. Look at number of patches and software changes that we have versus what they have got. It is a very different picture because Asterisk has been open source for so long and has developed such a community behind it. It is difficult to catch up as long as we do a reasonable job. OpenPBX is basically a variant of Asterisk. Actually it is a little bit confusing because there are two programs called OpenPBX. There is one program that Voicetronics makes and there is another one which is a variant of Asterisk.

What next in Asterisk platform?

Ans: Lot of the Asterisk features and development track is based upon developers in real world who tend to contribute features that are very specific to them. We are looking at doing more with standard based security. We are looking at doing more with standard based security to add TLS and SRTP support to Asterisk. There are other things like IAX encryption. Wide band and video are interesting areas as well as some more Jabber integration. These are the areas that seem interesting to us.