Voice in the Wilderness

The notion that a router and some data ports will provide the entire infrastructure required for VoIP to the desktop is simplistic at best. Until there is a universal workstation that can provide both voice and data capabilities and something that combines a PC and a telephone, we will need dual infrastructures.

Voice over IP sounds like a good idea until you ask - and answer - the hard questions.

I was trying to figure out why I was getting annoyed at my CIO friend Mark while he was making a valiant effort to convince me that voice over IP (VoIP) was the way to go to the desktop. I could understand the technical merits of VoIP, but I was still not convinced. And then it dawned on me: the conversation reminded me of the embarrassing questions a general manager asked when considering uninterruptible power supply (UPS).

A few years ago I was present at a meeting where a sales team was trying to persuade a crusty general manager at a manufacturing plant to upgrade his backup power. The upgrade would support six hours of uninterrupted computer operations - an improvement over the two hours the plant currently had. The sales team went on extolling the technical virtues of their new UPS system and the cost benefits the plant would get from it, when the GM, after listening politely for about 15 minutes, started asking the following embarrassing questions.

GM: Tell me Jim [plant IT manager], how long does it take to properly wind down our data centre after a power failure?

Jim: Oh, about 15 minutes; 45 if we want to run a backup before shutdown.

GM: Thanks. Do you guys [the sales team] know what happens to the plant when power is down for more than two hours?

Sales Team: Uhmmmm . . .

GM: If we have a power failure for more than two hours, we shut the plant down and send 99 per cent of the people home. That's because it takes more than three hours to restart the production line after power comes back on, and that requires about five people and no computers other than the ones in the data centre. Why do I need to spend almost a month's profit to keep my computers up and running while the plant is closed?

The sales team's answer? Silence.

Needless to say the sales team did not close the sale.

VoIP is one of those great technologies that people want to apply everywhere just because they can. And this brings me back to Mark. He was absolutely convinced that if he did not deploy VoIP to the desktop, his company would be at a competitive disadvantage. We decided to spend some time asking and answering embarrassing questions before Mark went to the operating executives with his request.

Question: What is the impact of moving voice traffic on the internal network?

Answer: Since voice traffic has different characteristics from data traffic and the human ear has low tolerance for people sounding like Donald Duck, we agreed that quality of service (QoS) and priority for voice traffic is important. But when we attempted to understand QoS features touted by the VoIP vendors, we ended up with very few assurances that they would work in an internal network environment. Bottom line: We don't know.

Question: How would changes made to VoIP affect daily telecom operations?

Answer: Mark's company experiences a move, add and change (MAC) rate of 35 per cent in its existing PBX phone system. (Each year, something is changed in 35 per cent of the phones.) Half of the MAC activity is physically relocating station equipment or new construction. The rest is software-related, such as changing routing features, display names, call coverage, phone numbers and so on. The people managing the PBX have this down to a science. A user simply calls or sends e-mail to the help desk with her request. Right now, the service-level agreement calls for software changes within four hours and physical changes within 24 hours. Unfortunately, there are no procedures for handling MAC activity in a VoIP environment. Bottom line: We don't know.

Question: Once VoIP is implemented, what new devices will the users have on their desks?

Answer: The notion that a router and some data ports will provide the entire infrastructure required for VoIP to the desktop is simplistic at best. Until there is a universal workstation that can provide both voice and data capabilities and something that combines a PC and a telephone, we will need dual infrastructures. Users will continue to have both a computer workstation and a telephone workstation on their desks, whether regular or VoIP. Since we are not sure what the network impact will be, we will probably have two separate networks to support VoIP. In effect, we would have to replace the PBX infrastructure that is simple, in place and paid for, with VoIP cabling and hardware. Bottom line: We'll need two of everything.

Question: What's the total cost of ownership (TCO) for VoIP?

Answer: That is the CFO's favourite question. We found plenty of cost analysis models for the backbone network but very little information for the desktop. Even something as simple as electricity to power the phones was not included in any of the models we looked at. Using the conservative estimates of Paul Rodecki, a Florida-based telephony industry consultant, the cost of powering 2500 VoIP phones for a year will be around $US65,000. Today's cost of powering telephone sets, with the exception of powering the PBX itself, is nothing. Another TCO issue that confuses many executives is long distance. Mark's cost for long-distance transport is less than 8 cents per minute (US) and dropping. Since his company does not have a large private network connecting its locations, calls have to be routed through a carrier's network. So implementing VoIP will not save him any money. Bottom line: We don't know, but the TCO for VoIP looks higher than that of today's phones.

Question: Are there any existing features in the current telephony environment that will be lost by going to a VoIP configuration?

Answer: Call-detail recording is the first thing that comes to mind. Mark's company has a strong cost accounting methodology for controlling telephony cost. How will call records be kept in a strictly VoIP environment?

Another major concern is the company's call centre. Many industry analysts talk about the flexibility of using VoIP telephones in a call centre environment, and I agree from a technical perspective. However, we could not find any applications to support call centre integration to the CRM initiatives Mark's company recently initiated. Bottom line: We will lose some features that we currently use.

Question: Are there any existing corporate obligations that would make adoption of VoIP difficult?

Answer: Industry consultant Tom Brophy from New Jersey-based NetPlus points out that most corporations have long-term agreements with their long-distance carriers. That's the case with Mark's company. Bottom line: There are some contractual issues that we don't know how to resolve.

Clearly, there is too much we don't know about VoIP. Given all the unanswered questions, why deploy a new technology to replace an existing one as reliable and critical as the telephone? VoIP to the desktop just isn't ready for prime time.

Chuck Papageorgiou is president and CEO of Discrete Wireless, a wireless ASP based in Atlanta, and is the managing partner of Ideasphere, a high-tech consulting company.

Join the newsletter!

Error: Please check your email address.

More about HISNetPlus

Show Comments