For the technologists out there that may be faint of budget, but not faint of heart, the open source movement is rapidly but quietly providing a cost-effective alternative to the commercial and often proprietary solutions for IP Telephony used by the vast majority of the world. I mentioned the OpenH323 project and related endpoint software OpenPhone in a tip a few months ago as a good tool to use for education, but as I've found over the past few months, there is much more out there than I expected, and I've found many more uses for it.
Education is still a fantastic way to take advantage of these free programs. Chasing corporate training budgets is like trying to find the pot of gold at the end of the rainbow. You sign up for classes and they just seem to evaporate. However, it's hard to deny free. Plus, the open source software can be downloaded anywhere, anytime. You don't need an expensive lab; a pair of old laptops is often sufficient. And if you have engineers all over the country, you're not forced to place the lab in a single place and fly your team around. It still takes some effort, but from reading the HOWTOs, RFCs and playing with the software, you can learn just about everything you want to know about the protocols.
Another excellent application is a test lab. Need to simulate a number of calls? Or do a proof-of-concept to show that your scheme for laying out gateways, gatekeepers and proxies is sound? Want to set up a standards compliant network so you can do interoperability testing on lots of different IP handsets and soft phones before you deploy them to your customers? The fact that software is infinitely more portable than a big stack of appliances, and doesn't require all the power, air conditioning, rack space and other accoutrements of hardware-based labs, means that you're much more flexible. This includes easy upgrades and the ability to rapidly change configurations.
Using free software in a proof-of-concept engagement has the added advantage of not requiring a ton of capital before you know it's going to work. Some commercial software, such as SIP Proxy servers can be quite expensive. Using a free version temporarily gives you the option of minimizing your loss if it turns out VOIP just isn't appropriate for your environment. And less risk means the project is more likely to be approved in the first place.
Of course, these open source programs can also be part of your "production" network. While some are clearly in pre-1.0 form, others have enough history to have a couple major releases worth of revisions and a respectable user-base.
First, the OpenH323 Gatekeeper, available at www.GNUgk.org, is now in version 2.0. Its list of features now includes support for gatekeeper clustering, call detail reporting, and MD5 and SHA authentication. With OpenH323 clients (like the aforementioned OpenPhone) and the GNU Gatekeeper (and a lot of patience and elbow-grease), your organization can have an internal IP Telephony network for free. If you decide to try the OpenH323 Gatekeeper, you should also test its web-based front-end, the Open IP PBX.
For connecting to the rest of the world, there are a few options. A number of free packages provide H.323 to ISDN gateway functions for WAN connectivity. These include the OpenISDNGw (http://www.gae.ucm.es/~openisdngw/home_en.php) and isdn2h323 (http://www.telos.de/linux/H323) (non-English) packages. For connecting to the PSTN, you can use the PSTNGw program available at http://www.openh323.org.
If H.323 isn't your protocol of choice, there is also plenty of software in the works for SIP. Much of this can be had at the Vovida.org website, which curiously, is sponsored by none other than Cisco Systems. Their products include SIP endpoints (soft phones), proxies, protocol translators for H.323 and MGCP and much more. This site also contains a link the some proprietary source code of a G.729 codec, made available for education and development purposes. That may be more than the average administrator wants to see, but it's good to know it's there.
Then there's the Bayonne project (http://www.gnu.org/software/bayonne/), which provides similar functions to a 'key system'. And certainly not least, the Asterisk* PBX software (http://www.asterisk.org/documentation_files/Asterisk.pdf), which has been well received.
Thomas Alexander Lancaster IV is a consultant and author with over ten years experience in the networking industry, focused on Internet infrastructure.