Taking a stab at open source IP Telephony

It’s been some time now that I was curious about open source IPTEL. Known in many names but all comes back to the Asterisk mother ship. Since it’s an open source application, it can save the potential customer some money. 

I started by trying to find the most simple way to get a filing for the application, without dual booting my laptop to Linux, and without installing and patching a Linux version and then installing Asterisk on top of it. VMware allowed me to run the software as a guest operating system and the simplified distributions of Asterisk did the rest. 

My first stop was AsteriskNOW, as published by its web site: 

Asterisk® in minutes. AsteriskNOW is an open source Software Appliance; a customized Linux distribution that includes Asterisk (the leading open source telephony engine and tool kit), the AsteriskGUI™, and all other software needed for an Asterisk system. AsteriskNOW is easy to install, and offers flexibility, functionality and features not available in advanced, high-cost proprietary business systems. 

It was a simple install indeed, and after some reading and configuration attempts, I was able to get a SIP phone to register and achieve basic functionality, but the web interface was not very responsive with Firefox (had to use IE) and I was not very happy with the available options. Since others had suggested trying TrixBox, I moved on and looked for this option. 

The trixbox install was easy (in VMware of course), the web interface was a lot better, I think this one is based on freepbx, and my sip softphone registered and allowed calls but there were still glitches in the interface so I decided to try another flavor, called elastix. 

Elastix interface was the best of all three and I had many options and stable GUI to work with, but I couldn’t get a decent audio quality from the integrated voice mail system. Later it turned to be that VMware requires a specific kernel patch, which I applied, but it still sounded bad. 

At the end I found myself going back to AsteriskNOW since this one was the only one that had a decent audio quality in VMware. 

After playing with those systems for a few weeks, my thoughts are that they might have a potential to do great things in the future, but right now an organization that has no exceptional technical expertise in Linux and Asterisk should take caution before committing to them. The reason is that the savings you will gain from getting this system are going to be swallowed by the additional expenses that will come from deploying it and getting what you want from it. The savings are only related to the price of the Call Manager piece of the system, you will still need phones, network switches with POE and gateways. You will also need a server to run the Linux machine on.

For a system that can support up to 50 users, you are looking at 3-5 k worth of savings. A price which is easily consumed by the time it takes to get this system to work. Also, if you are a GUI person, there are going to be text files and long CLI command that will be required. 

So why is it so appealing? I’m probably not the first one that figured out it’s not saving a lot of money. The main thing imho is that you can program your own code and or modify the existing one. If you are a gifted programmer, there is no limit to the things that you can do. Unfortunately that is not the kind of thing an enterprise CIO wants to build upon.  

Don’t get me wrong here, I’m seriously considering getting some sip phones and trying this system at home, it is a good thing to play with and we can always use our cells, but for a business, not so sure I’ll go that path.  

What do you think? Am I missing the point?

• Cisco
• VoIP / Convergence
• Asterisk
• elastix
• IP PBX
• open source
• open source IPP BX
• Trixbox