In order to understand a little better why I'm not a big fan of voice over IP (Vonage and the like) as a replacement for traditional telephone services in a business, it helps to know something about the initial design and development of the Internet. The development of the IP network (initially, ARPAnet, and now, the Internet) basically resulted from attempts to cure deficiencies in the telephone network.

It kind of worked like this:

The North American telephone network is built like a pyramid. At the bottom are all the Class 5 (local) offices; for example, the Monroe central office is a Class 5 switch. Above them (in ever decreasing numbers) are the Class 4 (Toll) offices. The Wayne, Michigan tandem exchange is a class 4 office. Above them them are the Class 3s (Primary), Class 2s (Secondary) and at the very top of the pyramid are the Class 1s (regional).

Calls through the network go up and down the pyramid. Say I want to call a friend in California. I dial 1-949-nnx-abcd. My class 5 office recognizes that it can't complete the call so it hands it off to the class 4 who can't complete it, so it hands it off to a class 3, etc.

The call might or might not make it up to a Class 1 office but at some point it will hit an office that knows what to do with it and the call will start moving down the pyramid, finally ending in a Class 5 office in Orange County.

That office will recognize the number as one of their own and send it off to my friend's phone. The entire connection, in all it's forms and permutations (analog, digital, copper, fiber optic, microwave, satellite) is mine for the entire duration of the call. That circuit is nailed all the way through.

The whole dialing process takes almost no time to complete and the connection quality is, usually, exceptional in quality.

So what is the problem? Why mess with success?

From the military's point of view (and remember, that's who really started the Internet) the system was flawed. Because there are so few Class 3, 2 & 1 offices, a failure of one or two of them could severely impact the ability of the military to communicate in times of war.

It was assumed that the onset of WW III would begin with the explosion of a nuke in the atmosphere to knock out all radio communications and then the sneak bombings of the Class 3, 2 & 1 offices in the US.

You could then call around the corner for pizza but the White House couldn't reach Norad.

Recognizing that the voice network, while not perfect was still excellent and not wanting to mess with it, the DoD decided that the new kid on the block, the data network, would get a complete makeover.

Recognize, that at that time data communications meant, for the most part, telex and TWX which worked with a network very similar to that of the voice network. Fast delivery wasn't as important as guaranteed deliverability.

So the idea was to develop a connectionless network. One wherein, if some of the major nodes were knocked out then the system would be self-healing and self repairing, and would have the ability to route around the damaged nodes transparently - and all the data traffic would get through (even if voice was blocked). The goal was that, even if the shortest path is operational, the system would force later packets into alternate routes - just to maintain diversity.

And so the Internet was born.

It works incredibly well for data traffic. But data, while it requires much greater bandwidth than voice does not require immediacy. The Internet by design cannot deliver (or at least guarantee) immediacy.

So, my position is that, while VoIP is fun to play with, and offers some real cost savings in certain applications, it will never work well (or at least as well as we'd like it to) over the Internet because the Internet was specifically designed not to handle it.

Now, having said that... I have a VoIP phone at home, and I use some VoIP trunks in our office.

I'd just never be comfortable relying entirely on VoIP in a mission critical application like business telephone service.