TORONTO (03/16/2000) - As any orator will tell you, if you want to communicate to the best of your ability you need a good set of pipes. That's why more and more organizations are considering converged networks that simultaneously carry voice, data and maybe even video. They are doing so for two main reasons: the potential cost savings such converged networks promise and the new applications they could make possible. Helping the process along is "the near-universal acceptance of TCP/IP," adds Ross Chevalier, director of technology at Novell Canada Inc. There is no longer much question what the common foundation for converged networks should be: It is IP -- Internet Protocol. Yet at this point in time few are doing much more than looking, or at most bringing voice and data together in limited areas of their operations. Why? First, it remains doubtful how real the promised cost savings are. Avoiding long-distance charges accounts for a significant chunk of the savings that convergence advocates promise, but with increased competition driving long-distance rates down, some doubt there is a lasting gain here. And while combining two separate networks into one should cut management costs, those savings will come only after the pain and expense of reorganizing and retraining the IS and telecom teams that manage existing networks. New applications may eventually prove the more convincing argument for convergence. Unified messaging -- in which users can retrieve voice mail, electronic mail and faxes through one interface -- is one.
Another is the integrated customer-service centre, which expands the call centre's role to dealing with customers online and even carrying on voice conversations through Web sites.
Show Me The Savings
Ideas like these may be tempting enough to get organizations moving over the next few years, but convergence would probably happen faster if managers were more certain of cost savings. "There are few things more powerful than saying 'Mr. Customer, I can save you money here,'" notes Dan McLean, an analyst at research firm International Data Corp. (Canada) Ltd. in Toronto. So, some vendors are trying to make that case. For instance, Andrew Sage, manager of Canadian marketing at Cisco Systems Canada Ltd. in Toronto, says Cisco's studies show that a company operating one or more branch offices with a total of 100 users can see a return on investment of 169 per cent in three years by combining voice and data traffic. According to Sage, 23 per cent of the savings would come from avoiding long-distance charges. Another 30 per cent would come from staff reductions and 10 per cent from other savings on maintenance. The largest chunk of the cost savings, though -- 37 per cent -- would result from cutting the cost of network moves, adds and changes. Sage says a converged network makes these much simpler -- to move an IP telephone, for instance, just unplug it from its old location and plug it in at the new one. Sage acknowledges the picture is rosiest for organizations with many locations, because the greatest savings potential lies in longer-haul links. A converged network within a single building or small campus can bring some savings in maintenance and support costs, but the toll bypass component is missing. For an organization with 800 of 1,000 employees in one central location, Cisco says, the return on investment in a converged network would be about 136 per cent. Combining voice and data will not eliminate the entire telecom or data communications budget, Sage says, but might save about a quarter of the total of both budgets. Some are less optimistic. Jean-Pierre Levesque, Eastern region vice-president of sales for network integrator Charon Systems in Montreal, says his company stopped installing integrated voice-data connections using voice over Frame Relay on top of IP because the installation cost was too high.
Levesque says the IP private branch exchange (IP-PBX) products on the market today lack many functions of conventional PBXs, and are difficult to cost-justify. And with long-distance rates dropping to around the five-cent-per-minute mark, he says, reducing long-distance bills isn't as compelling as it used to be. McLean agrees the savings are hard to prove, and he notes that Cisco is one of very few vendors pushing cost-cutting as a selling point for convergence. Others, including Nortel, Lucent Technologies Inc. and Compagnie Financiere Alcatel of Paris, are focusing on the flexibility of IP-PBX products built on standard computer hardware. Brian Allain, vice-president of Internet business systems at Lucent, says savings vary widely depending on the customer's existing networks and long-distance rates. He does acknowledge, though, that "the cost difference is probably going to decrease over time." The emphasis in selling converged networks will be less on cost savings and more on other factors in the future, Allain says. Cost savings "is certainly part of the story," says Ric Walford, director of network consulting at 3Com Canada Inc. in Toronto, but "maybe not the largest part." Walford prefers to talk about a "whole group of applications coming to market that can take advantage of convergence."
New Wave of Applications
To date, Allain explains, computer-telephony integration has usually required a separate server that integrates with the PBX. An IP PBX, he says, will offer application program interfaces (APIs) so customers can build new applications directly on the same server. "The net effect of this is that the cost of new applications is going to be going down," he predicts. That in turn will mean applications that could not be cost-justified in the past will become viable.
One application Walford, Sage and others dwell on is the converged call centre.
It is already possible today to link voice and data systems so that when a call comes in, that customer's record pops up on the appropriate representative's screen. "But it's quite a bit of work to make those systems interact," Walford says. An integrated network would make it much simpler. Sage says new possibilities will open up, such as intelligent contact management applications using the Java programming language, and the ability to add on customer-service staff working from home during peak periods. Allain adds that integrated networks will make it easier for small businesses to build "informal call centres," offering many advantages of a centralized call centre but not forcing employees who take the calls to sit together in one place and do nothing else.
Unified messaging is also capturing a good deal of attention. With a single network, it will be easier to present voice messages, e-mail and even faxes in one multipurpose mailbox. That will mean travellers can have their e-mail read to them through any telephone, while users will also be able to click on the voice messages they want to hear, rather than picking their way through multiple messages using a telephone keypad. Other possibilities include better collaboration capabilities, in which co-workers in different locations could look at the same documents while carrying on a phone conversation.
Today, network convergence is primarily about putting voice and data through the same pipe. However, there is a third type of traffic in the backs of many network managers' minds. Video is used relatively little in the workplace today, but with high-bandwidth networks it will become more attractive.
Technically, Sage says, video over converged networks is as feasible as voice.
Because of the bandwidth demands, though, it may take longer to see widespread use. The other issue, adds Lucent's Allain, is that users are not as accustomed to it yet and thus the demand is limited. The possibilities include videoconferencing and, perhaps even more interesting, video-based training delivered right to employees' desks. Distance learning is an obvious application for video, and many in the education field, such as the Education Network of Ontario, are interested -- but distance learning is not limited to schools and universities. Corporations have a continuing need to train employees, and this is possibly the single application of network video that excites the most people. "That's where the real demand is going to be," says Novell's Chevalier. As the idea of video becomes more familiar, other applications not yet thought of may arise too. "Once you have superconnectivity to the desktop," observes Tom Hope, senior vice-president of technology and operations at Bell Nexxia, "then you're going to communicate in very different ways."
When Worlds Collide
Such promises are all very well, but converged networks are bound to bring some headaches too, especially for those who must keep them working. Levesque at Charon Systems says one of the greatest obstacles to convergence today is a shortage of people with the skills to build and manage these integrated networks. Traditionally voice and data communications have been separate camps, and few people understand both. Worse, the two groups speak different languages. For instance, Levesque says, the term "bridge" is familiar to both telecom and datacom people, but means something different to each group. So anyone who tries to solve the skills problem by bringing together voice and data experts risks creating a tower of babel. The other major stumbling block is the ability to deliver high-quality voice transmission over an IP network.
The problem is not that it cannot be done; it can. However, the network must be up to the task. The public Internet is not up to it, at least not yet.
Provided the network is private, or one built by a service provider with the quality of service requirements for voice in mind, voice over IP is feasible today. But even if the technological problem has been solved there is still a problem of perception. The perceptions have to do not only with the quality of service issue but with reliability. Many people perceive data networks as less reliable than voice networks -- most of us are used to office data networks being down fairly often, but we rarely pick up our telephones and find no dial tone. However, Walford argues, in reality these network problems are usually caused by servers and other devices on the networks. "It's very rare that the network itself is broken," he says. Whether the perceptions are entirely fair or not, they exist. Because of these and other problems, McLean at IDC says, "customers are still sort of in watch mode." However, there are strong forces working toward network convergence. One is the economics of the telephone business. The market for voice services is mature and fairly flat, McLean says.
Telephone carriers see great opportunities in providing data services to customers. For instance, Gateway Telecom Canada Inc. recently launched a network that officials of the North Bay, Ont., company said will help emerging competitive local exchange carriers provide both voice and data services at low cost. Carriers could save 15 per cent to 20 per cent of the cost of providing the services over distinct networks, and get services to market faster, says Tim Kuntz, Gateway's executive vice-president. The incumbents are determined not to be left behind. For example, Bell Canada's sister company, Bell Nexxia, is promoting wide-area links that will carry both voice and data as a way for its customers to save money. "Every carrier, if they haven't already built IP infrastructure, is in the midst of doing that," says McLean. The move to converged networks will be an evolution. Businesses will keep private branch exchanges until they are ready for replacement, and put in combined voice and data networks as they make business sense. "It's not like we're just going to turn off one and jump on the other," says Kuntz. Whenever and however we choose to embrace converged networks, one fact we can't ignore -- it seems quite likely that sooner or later, CIOs will have to deal with them.
Grant Buckler is a freelance writer specializing in information technology and IT management. He is based in Kingston, Ontario.
Sidebar: Solving the Latency Problem By Grant Buckler CIO Canada Transmitting voice traffic over a packet-switched network designed for data presents one major problem. When a data packet is delayed slightly on its way through the network, it makes little difference, but when a packet making up part of a conversation is delayed, the person at the receiving end hears the difference. The caller's voice seems to be breaking up. This issue of the latency (or delay) in delivering voice packets has always been one of the major objections to voice over Internet Protocol. In fact, this is more of a problem with transmissions over the public Internet than with traffic that flows over a private network. The public Internet is essentially unmanaged, notes Ross Chevalier, director of technology at Novell Canada Ltd. It is easier to manage the private network to ensure voice traffic gets through on time. Up to now, conventional wisdom has been that the way to do this is with Quality of Service, or QoS, capabilities in the network. In effect, voice packets are sent by special delivery to make sure they get there on time and in sequence. Many debates about the merits of different high-speed networking protocols centre on their ability to support this. As high-bandwidth optical networking technology becomes more widespread, though, some are proposing another way of dealing with the latency problem. Forget QoS capabilities -- just throw bandwidth at the problem. "If you have very high bandwidth," says Tom Hope, senior vice-president of technology and operations at Bell Nexxia, "then you really don't care so much." David Isenberg had been working at AT&T Bell Labs for about 15 years when, in 1997, he wrote a paper entitled "The Rise of the Stupid Network". His thesis was that instead of building more intelligence into networks, we should simply think of bandwidth as cheap and plentiful, and use it to solve network problems such as latency. The paper drew so much attention that Isenberg left AT&T to form his own telecom analysis firm, Isen.com of Westfield, N.J. Nonetheless, not everyone is a convert to Isenberg's viewpoint.
For instance, James Treleaven, at the Education Network of Ontario, says he doesn't believe throwing bandwidth at the problem will work. "What I do think will work," he says, "is QoS guarantees backed by the networking equipment and sold as a service by the IP carriers." The debate rages on.
Sidebar: Voice over IP in the Classroom By Grant Buckler CIO Canada When employees at the Education Network of Ontario use the phone, they're doing something unusual. Phones in the provincial agency's Toronto office work over an Ethernet local-area network. The Ethernet phones are one step in a plan to add voice capability to the province-wide Internet Protocol (IP) network that ENO -- a Ministry of Education and Training initiative -- has created to connect Ontario's schools. Eventually, says ENO sales engineer James Treleaven, similar phones may appear in Ontario's classrooms. Treleaven notes that schools today usually have only a couple of phones, because of the cost of running lines to every classroom. Yet many are wired for data, and if the same networks could be used for phone service, teachers could have more accessible telephones. There are other possibilities. Employees at ENO's Toronto office already can use collaboration tools to share documents on computer screens as they discuss them by telephone. Since they are close together, they don't use this capability much yet, Treleaven admits, but when it becomes available across the province, it will be much more interesting. Eventually, the network will also handle video for distance education and videoconferencing. However, it will not happen quickly. Many schools rely on dial-up access to the ENO network, which has 28 points of presence. Before they can enjoy the full potential of a single network for data, voice and video, those schools will need dedicated links.