Testing times for voice and data traffic

Voice and data are converging, but perhaps not as quickly as everyone seems to think.

The combined market for carrier-type dial-up internet access and voice over IP systems was $3.7bn (£2.4bn) in 1999, according to analyst IDC. It predicts that the market will grow to $6.7bn by 2003. Researcher Cahners In-Stat Group predicts that worldwide voice over IP gateway sales alone will reach $3.7bn by the end of 2003.

Last year, a Forrester Research survey of 52 Fortune 1000 companies in the US concluded that more than 40 per cent of telecommunications managers plan to move some voice or fax traffic to the internet by the end of 1999. IDC has also said internet-enabled communications could account for about 11 per cent of US and international long-distance voice traffic by 2002.

Two become one
Voice and data convergence is happening, at the applications level in particular, and the products that enable it to happen are ready and waiting. However, they are not always being used to converge the traffic at a physical layer, because there's simply too much of a legacy in private branch exchange (PBX) systems and no vendor wants to be the first.

"Many traditional communications managers don't like taking risks," says Paul Templeton, director of portfolio and solutions marketing at Nortel. "Most analysts predict that something like 30 to 40 per cent of voice systems installed over the next five years will be based on some kind of call centre application. In other words, they will be converged solutions. That tells you that for some time you will have mixed environments."

Recent debates on this issue always reach the same conclusion: convergence of voice and data is possible now, but most of the time it happens at the applications level. The physical integration of voice and data is not happening as quickly because the costs and disruption are too high, and those holding the purse strings regard it as too great a risk to take with something as fundamental as voice services.

Many more organisations are installing systems capable of carrying both voice and data now, but initially are using them for either one or the other service, says Rob Sims, UK business development manager at Swedish carrier company Telia.

"We have seen rising interest in the past six months and we are now seeing deals come through, especially for direct connections. A single pipe can carry voice, data and IP traffic, but they do tend to use it for specific purposes. They are only just starting to do this for voice and data," he says.

Don't believe the hype
Nor is voice-over-IP happening as quickly as the publicity delivered by some vendors might suggest. The key difficulty with it remains its quality. Putting voice across the Lan or even a Wan with fixed links is fine, says Templeton, as you have control of all the points on the network. But once you go outside the network, on to the internet, control is lost.

This is the reason service providers such as Telia are seeing demand for direct connection services, and also why activity in this market is, for the moment, mainly concentrated within the M25 and major cities. London and other urban centres are the first places where the fibre, cable and carrier companies have located their networks.

In this region, voice resellers in particular are being drawn into the market, says Sims. "Once they have grasped the benefits they move very quickly. Well over 60 per cent are showing a positive interest now and have business going through or in the pipeline."

They move fast, says Sims, because the advantages of providing both the data and voice part of the solution quickly become apparent. "You've got lock-in, you are taking all the traffic of that customer and it also gives you a better perceived image. You become more than just a reseller, you are providing something different, and it takes you to a higher level in the business."

The main attraction
Voice and data is not just an IT decision. Other senior managers in the business want to know about its savings and the effect it will have on voice communications. It is this promise of lower costs that attracts companies to voice-over-IP, and the involvement of other managers that holds back adoption.

The quality issue in particular is a problem. Stuart Muirhead, director of sales and marketing at Trend Network Services, says that even with the current version of IP, IP4, there are serious limitations in the addressing range of voice-over-IP. He feels greater bandwidth will be needed to deliver adequate services.

"The technology is not yet mature enough to address the quality of service issues, although this will happen over time. Voice-over-IP is not likely to become a primary point of communication, rather it will become a reserve option. Voice-over-frame is likely to have more impact."

It is a view echoed by others and one that points to more direct connection structures being used for converged voice and data, rather than fabrics that switch traffic over the public internet. The quality of service may never be good enough to carry voice through that route on a day-to-day basis.

Marcus Ashby, a senior account manager at internet consultancy iPcenta, says the key to making a voice-over-IP network perform is to work with a single vendor that has very high quality of service standards.

There has been dramatic improvement in performance in key areas such as RSVP, priority queuing and latency, and has refined the process of checking and reserving the bandwidth needed to handle voice traffic over a network designed to carry data. Ashby also points out that more ISPs are developing peering arrangements and contingency plans for traffic routing and bandwidth allocation.

While these improvements are not convincing managers overnight, they are persuading more to look at the technology, says Ashby. "Companies can still maintain their current PBX systems and integrate them slowly with the internet on a pilot basis, one office at a time, which is also pushing things forward."

Counting the cost
The arguments for convergence are too strong for larger organisations in particular to ignore. They include large savings on long-distance telephone calls, call logging and more integration and control. But adoption is always for a specific purpose, or otherwise is applications-driven.

Office politics can also hold back adoption of a converged solution, says Ashby. Unusually, it is often senior financial staff pressing for convergence, rather than IT or communications managers.

"It is a very interdepartmental decision-making process and also one that requires very new skillsets. For the first time we see the financial director of a company pushing the IT department to look into voice-over-IP because of the heavily advertised savings. The IT department is then faced with the prospect of working out how to connect traditional telephony with the internet, and to many this a very daunting task," he added.

Wait and see
But many businesses may do nothing until they are compelled to act. The generally accepted wisdom in the industry has been that major companies would be the earliest adopters. But Chris Hornsey, UK sales director at communications systems supplier General DataComm, says the market won't be driven by the business-to-business world alone.

"Residential users are now expecting the substance behind the communications revolution hype. They've been sold the idea of interactive TV, radio, video on demand and omnipresent, real-time internet. They will expect to get the same functionality at work, dealing with their business partners and from their own IT infrastructure."

The first place to sell converged solutions, he says, is at the new-entrant level and to organisations addressing the retail market, because this is where expectations will be highest as the market develops.

Mark Randall, managing director at ATL Networks, a distributor specialising in IP and managed services, says that while cost reduction has driven voice-over-IP adoption to date, with deployments usually limited to internal corporate networks or so-called calling card applications, we are very close to seeing major new breakthroughs in this area.

"There have been some limitations and quality is a concern. But the technology is there to deliver voice-over-IP. It's just that none of the ISPs or the networking companies have yet delivered a commercial package. We are committed to driving combined data, voice and application services through the UK channel, and have invested a lot of time and money in this area over recent months," he says.

Randall points out that to date we have witnessed only the first stages of the internet revolution. There has been an initial dash for numbers of subscribers by ISPs, and the recent flurry of activity and investment in dotcom companies has been based largely on business-to-consumer ebusinesses.

"When the euphoria dies down and people look at how they can really benefit from the internet and from IP, saving money by using it to carry voice traffic is one of the first areas they will look at. But to do that, ISPs will need to deliver a very special kind of service, and most don't have the capability of doing that by themselves. They are not competent at delivering applications themselves. If they were, the concept of an application service provider (ASP) would not exist," he says.

Turning the key
Just as there are specialist business service ISPs such as BT and INS now, Randall predicts the emergence of providers that will provide integrated voice, data and application services. Such services are not likely to be delivered by ISPs or ASPs, but consolidators such as ATL that can select best-of-breed components and create integrated 'turnkey' solutions.

"These services will not be 'me-too' services. They will be highly specialised and will address the business-to-business market with rapid return on investment. They will need to be properly configured and managed at the customer end," says Randall.

The key applications areas are already emerging that will drive the business market in this direction. Call centres, unified messaging, web access management and ecommerce that combines a voice service with online buying are currently the main areas of focus.

The convergence market is all about applications, not technologies, says Tony Duerinck, marketing manager of Dialogic Telecom Europe, an Intel subsidiary that provides computer-telephony integration products. What we should not expect is for voice and data to be a revelation for the user. But creating new applications by combining web browser control and management with language speech recognition and unified messaging technology is a different story, he says.

The key points

• applications are driving the adoption of converged voice and data solutions
• mixed public switched telephone network and IP-based voice solutions are being adopted by many companies as they seek to take advantage of the technology
• it is still voice or data rather than voice and data that users are installing as they put systems in to solve specific issues or provide services. Data and voice being carried on the same wires is still an additional benefit
• quality of service remains a major restraining factor to wholesale adoption of converged voice and data technologies
• new and complete services that deliver the whole converged voice and data solution are needed to drive the market forward.