Whether voice over IP is good or great is a matter of debate. But in some respects, VoIP is like greatness.

In Shakespeare’s Twelfth Night, the character Malvolio said: “”Be not afraid of greatness: Some are born with it, some achieve it and some have greatness thrust upon them.””

In another

decade, some will have been born with VoIP. Some early adopters have achieved VoIP, overcoming the difficulties posed by quality of service, network management problems and the challenges of bringing telephony services such as voicemail, call forwarding and emergency 911 services to the IP world.

But most tech services staff over the next decade will, rather than achieve VoIP implementations, simply have it thrust upon them.

Today, most organizations have the choice of purchasing telephony products based either on IP or on time-division multiplexing (TDM) technology. But in some respects, the decision to buy VoIP tomorrow will be like the decision 10 years ago to either buy Microsoft Office, or stick with a combination of WordPerfect 5.1 and Lotus 1-2-3. The question of whether to move to an all-IP network will be based on what everyone else is using, what the vendors support now and what they will support a few years from now.

One of the major deciding factors in buying VoIP has been return on investment (ROI). The major IP telephony vendors tend to market VoIP in the enterprise on the basis of the applications it enables, rather than the cost savings. This makes sense in North America, where long-distance rates are heading in the opposite direction of crude oil prices and few can justify an IP telephony purchase by the cost savings achieved by bypassing the long-distance networks.

Startup costs vary widely

Several market research firms have done ROI analyses on voice over IP. For example, New York-based Nemertes Research LLC published a study last November that found startup costs range from US$515 to $1512 per user, depending on which vendor’s equipment is used, and the total number of seats installed. But calculating cost is not as simple as adding up the price of the hardware components and an integrator’s labour and consulting fee. It also includes planning before the implementation, and troubleshooting afterwards. Nemertes’ figure includes a “”baseline network assessment,”” which costs an average of US$17,220.

Although this assessment costs about as much as it does to hire a senior IT person for a few months, it can be difficult to justify to a senior executive who wants to know how much new revenue VoIP will bring in or how much money it will save. But basing a VoIP purchase decision on ROI is like a person basing their decision to buy a car on ROI. A cost-benefit analysis has some use, but sometimes it misses the point.

Suppose you live near and work near major public transportation routes, and you can take the train and/or bus to work in about the same amount of time it takes to drive. If you don’t need to get to any place that’s not accessible by transit, you may not need a car. Add up the cost of public transit and cab fares, compare it to the cost of buying a car, maintaining, fueling and insuring it, and public transit is less expensive.

But most people are not going to go without a car just to save money. You may choose not to own a car simply because you don’t really need one. If your circumstances were different, you might buy one, and if you did, the ROI would not be measured in dollars and cents. Instead, the purchase would be based on the fact that the alternative — not being able to get to work — would be unacceptable.

This is a very long-winded way of advising IT and telecom managers to purchase voice IP if — but only if — they need it. This sounds pretty obvious, but what’s not so obvious is the difference between what you would like to have — what looks and sounds cool — and what your organization actually needs.

The cost of maintenance

Much like a new car, voice over IP looks and sounds cool if you see it on a trade show floor or in a manufacturer’s demonstration environment. When used properly on a network that can handle voice traffic, it’s quite jammy in a real-life implementation.

But the network, like the car, needs tuning and maintenance. Assessing and monitoring networks has become a cottage industry over the last few years, with firms such as Viola Networks of Tel Aviv and Apparent Networks of Vancouver specializing in tools designed to figure out whether a network designed for data can actually support voice. Although the problems of voice quality and the underlying network have largely been addressed, there are still technical problems that can make the Lambourghini Countach of IP telephony products run like a Trabant from the former East Germany (for more information, please see Getting the Packet Deal, page 14).

As carriers and service providers start to offer managed IP telephony — which would let enterprises run VoIP without having to buy IP PBXs — companies can use IP telephony without having to worry about choosing one particular brand of IP PBX. They only need to worry about the end points.

Several questions remain: What can I do with voice over IP that I can’t do now? Do I need all the features? If my competitors buy this, what will they be able to do that I cannot do now? Until we have VoIP thrust upon us like Microsoft Office, these are questions telecom managers should be asking their vendors.

Share on LinkedIn Share with Google+