ITEXPO begins in:   New Coverage :  Asterisk  |  Fax Software  |  SIP Phones  |  Small Cells


Open Source Telephony Is Cheap to Buy, but Can Be Costly to Manage

By David Mandelstam


Anyone who has had the experience of installing a proprietary PBX or key system will know firsthand how the system itself � expensive as it is � can represent simply a down payment for a long and costly ownership experience. In the case of Open Source telephony, where initial costs are much lower, the ratio between first costs and total cost of ownership can be even more marked and can be the source of considerable frustration.

For any business, the telephone system still represents the most important contact point to the outside world. Notwithstanding the rise of e-mail, messenger services, even cell phones, the main contact portal for any company still remains the telephone system connected one way or the other to the PSTN. For some business, like call centers, the system being down results in immediate, quantifiable, and crippling losses, but even for less telephone-centric businesses, loss of PSTN access, even for a few minutes, is traumatic. So, it is important to realize that often the biggest hourly expense related to a PBX of any kind is the cost of down time. A billing rate of $300 per hour looks like a bargain when your company is entirely without phone service.

This is something to think about before deciding to build your company PBX using free, community supported beta software on a combination of the cheapest PC you can find and semi-experimental hardware of questionable quality.

The open source PBX implementations, such as Asterisk, support features that would be beyond the reach of small businesses and, in fact, beyond the reach of almost anyone. They provide the flexibility to integrate voice and voice processing into data and operational systems. They provide an easy way to transition into the world of VoIP. But, in many cases, the prime objective has been the saving of money, in particular, first cost.

The cost savings relative to proprietary equipment is considerable. But, don�t be tempted to cut corners to extract maximum savings. Rather, treat the savings as a windfall and plan to invest a portion of the savings in system integrity.

For installation, set-up, and support, it is tempting to imagine that, because you have a Linux system administrator on hand, he can easily handle that PBX. After all, the soft PBX is only another Linux application among others and, once installed, the amount of work on a PBX should be fairly limited.

It�s true that any competent Linux system administrator is perfectly capable of installing and maintaining one of the open source PBX implementations. There are no special arcane skills required and the configuration files and setup scripts are arranged logically and are fairly easy to understand. The problem, however, is that anyone in charge of a single PBX is going to be working on the system sporadically and, inevitably, under pressure. Skills picked up twelve months ago are going to be rusty when dealing with today�s new crisis.

For businesses that are not primarily in the telephony business, it is probably good practice to make use of an outside consultant. A consultant is presumably looking after many systems and has sharp skills. She has also doubtless picked up some neat tricks in the field and has experience with which parts and versions of the open source telephony project work well and which are still experimental. The outside consultant is usually not very expensive because, unlike your old PBX consultant, she very seldom has to come on site. Almost everything she needs to do can be done through the magic of the Internet and SSH, from debugging driver problems, through setting up of dial plans and provisioning telephones.

What about hardware? Here again, a little investment goes a long way. Don�t buy the cheapest motherboard/PC combination you can buy. Use a server supplied by one of the �big three.� They are not that much more expensive, and at least you know that the different components have gone through a solid QC program and have been tested working together.

RAID is a pretty good idea. Disks are the least reliable component of any system and they do die. Backups are often not quite up to snuff when disaster strikes. RAID represents a small cost premium for considerable insurance.

If your PBX has a traditional PSTN access there will be additional hardware needed to provide the analog or digital interface. Choose this hardware with an eye to quality. If you are not careful, you can spend endless person-hours integrating PSTN hardware and getting it to work with your chosen system.

Remember that, as a general rule, hardware is less expensive than labor. If you can reduce person-hours by substituting hardware, it is almost always money well spent. For instance, hardware echo cancellation on a six-port FXO system at about $50 per port seems very expensive. But, if you compare it to the cost of several hours of system tuning needed to get software echo cancellation to work properly, it�s a bargain. And if the system grows, it is nice to know that you have no further costs for echo cancellation.

A popular alternative is to install a pre-packaged, pre-configured system based on one of the Open Source projects. These can save enormous headaches because they have been through a testing program as a complete package. The packages include expert help in setting up the system and usually the user interfaces have been engineered for easy self-configuration with minimal risks. The costs are higher than the components of a homegrown solution, but the TCO is incomparably better. The only drawback to these PBX systems is that inevitably the customization and pre-packaging limit the flexibility. However, there is nothing to stop you configuring by hand to add features as you gain confidence. After making good backups, of course!

Open Source telephony projects can provide spectacular bang for the buck when compared to commercial closed systems. But, you should make sure that the goal of reducing costs does not compromise system reliability. It is certainly possible to achieve commercial levels of reliability and uptime for a system based on solid, reliable, hardware and software and with a competent, responsive support team. IT

David Mandelstam is President of Sangoma Technologies (news - alerts).

Sponsored Editorial



Today @ TMC
Upcoming Events
ITEXPO West 2012
October 2- 5, 2012
The Austin Convention Center
Austin, Texas
The World's Premier Managed Services and Cloud Computing Event
Click for Dates and Locations
Mobility Tech Conference & Expo
October 3- 5, 2012
The Austin Convention Center
Austin, Texas
Cloud Communications Summit
October 3- 5, 2012
The Austin Convention Center
Austin, Texas