You could be forgiven for concluding that voice over IP (VoIP) is now a done and dusted deal - or, at least so in its wireline flavour. What started out as something of a low-quality hobbyist's curiosity has fairly rapidly grown into a major component of the long haul wholesale carrier business. According to data from the TeleGeography traffic analysis and research company, global voice traffic totalled nearly 200bn minutes in 2003, of which 22bn or 11% was carried over the Internet. Extrapolating from the same company's projections for 2005, global VoIP will this year exceed 40bn minutes, or over 14% of the 280bn wholesale total.

Now, as any number of analysts' forecasts tell it, VoIP is making a determined and seemingly successful assault on the Class 5 retail voice business. At the forefront of this assault is the corporate user community. According to a 2004 report from European industry analysts IDATE, whereas only 2.5% of business lines were operating over IP at the end of 2003 (10mn lines), by 2009 close to 40% of business lines will be IP-based (180mn lines), regardless of the solution used. IDATE calculates this will mean the disappearance of close to a third of classic analogue and ISDN lines.

Again, the bi-monthly 'Radicati Market Stats' published in March 2005 by The Radicati Group Inc had corporate IP telephony revenue reaching US$5.5bn by 2008, an increase from US$1bn worldwide in 2004.

First adopted by the very largest enterprises, VoIP is now percolating down into more and more small and medium businesses (SMBs). In the 2004 AMI-Partners' bi-annual worldwide SMB IT and telecommunications sizing study entitled 'The SMB Global Model', worldwide SMB VoIP telephony spending is forecast to grow from US$1.2bn at the end of 2004 to over US$4.5bn during 2008.

As VoIP is moving down the enterprise scale, it looks to be spreading out across the geography, particularly to Europe. In another recent bi-monthly 'Radicati Market Stats', it's estimated that in 2004 45% of the corporate IP telephony installed base was located in North America, followed by Europe (38%), Asia/Pacific (15%) and RoW (2%). By 2008, the analyst expects the installed base in Europe and North America to even out to 40% and 41% respectively. The Radicati Group further estimates that 90% of today's IP telephony lines are powered by hybrid PBX systems, but that pure IP-PBX systems will account for 30% of the total by 2008.

In tandem with the business community, and spurred on by the efforts of enterprises such as Skype and Vonage and others, consumers are switching to VoIP in ever increasing numbers. "VoIP has now also become reality for residential customers and network operators", observed Siemens Communications Group ceo Lothar Pauly on the eve of this year's CeBIT extravaganza. "The market for personal consumers, who are using VoIP at home, is booming: it will be growing at an average of 80% annually through 2009".

Luxembourg-based Skype Technologies SA claims to be the leading VoIP-category product worldwide, with more than 24mn registered users, and more than 130,000 new users joining per day. The Vonage VoIP vital statistics at February 2005 were more than 400,000 lines in service growing at over 30,000 lines per month, over 10mn calls per week, and a predicted 1mn lines in operation by the end of 2005.

One estimate from Nadahl Shocair, ceo of DeTeWe UK and directing strategist, DeTeWe AG & Co, is that voice over Internet revenues are going to be worth between US$135bn and US$180bn over the next three to four years, equivalent at the higher figure to 40% of current incumbent revenues.

So, end of the incumbent-and-circuit switched TDM voice story? Not quite.

A fairly widespread assumption is that VoIP is a particular threat to traditional incumbent telcos. A related assumption is that the chief service provider beneficiaries of VoIP will be the fleet-footed start-ups that are pioneering the technology. Shocair, for example, has characterised national operators as 'dinosaurs'. "The telecoms market is changing at a rate of knots. Disruptive technologies – namely VoIP - from providers like Vonage and Skype promise to batter the incumbents’ revenues at an alarming speed as IP telephony takes off", he contends. "As the voice and service quality of VoIP improves, the enterprise market – seen by many as the saviour for the incumbents in terms of keeping revenues alive – will also switch to IP – probably in the form of managed IP services. These are interesting times – unless of course you’re an operator relying on traditional circuit-switched voice revenues".

Can dinosaurs learn new tricks?
You betcha. The incumbent as VoIP loser/start-up as VoIP winner scenario may come true for particular service providers, but ignores a number of market realities. One is that some established operators will join forces with the start-ups: in February, for example, Hutchison Global in Hong Kong became the first fixed telecommunications network services (FTNS) operator in the world to reach a co-branding agreement with Skype.

At the same time quite a number of the so-called dinosaurs - BT, Telecom Italia, Telstra and SBC to name but some - are themselves hitting the VoIP trail big-time. A recent meeting of the European Regulators Group (ERG) considering VoIP regulation (click here) concluded that it was impossible at this stage to predict what might happen in this increasingly volatile market. However the ERG also noted: "…it has been argued that although new innovative providers could push the market, in the long run it would be the established players (including incumbents) with their own networks that continue to dominate the market due to their reputation, reliability, as well as their established customer base and ancillary services such as billing systems, and so on."

The ERG added: "The cost structure of IP networks with high fixed costs and low marginal costs of service provision may reinforce a trend towards flat rate tariffs for users. Such a trend may be advantageous for those providers with established customer relations. This advantage applies specifically to all providers controlling the access infrastructure (irrespective of the technology applied)".

The commercial prospects for VoIP start-ups will also be shaped by the consideration that peer group competition is predicted to be ferocious, and the expectation that voice service margins will be very thin. These sorts of factors may not be of such consequence to incumbents and longer established operators who have VoIP as one component of a mixed bag of converged services, notably the famous triple play deal. Nor are the economic benefits s of VoIP confined to the start-ups. A new report from the Analysys consultancy - 'The Business Case for Carrier Migration to VoIP’, for example, calculates that the migration of a Class 5 office to VoIP using existing copper for both voice and broadband could save European incumbent operators 11% of voice network costs.

Regulatory debates rumble on
Again, although no-one is really anticipating that regulation will now be an absolute VoIP show-stopper, what shape regulation eventually takes will obviously impact the business opportunity for start-ups and established service providers alike. The plain fact is that VoIP regulation is still evolving at different speeds and with somewhat different agendas in different parts of the world.

In Europe, for example, there has been some focus on the issue of Electronic Communication Services (ECSs) - of which VoIP is one - and Publicly Available Telephony Service (PATS), which is a subset of public ECS. Interalia, classification as a PATS obliges the service provider to furnish uninterrupted access to emergency services. One dilemma for the VoIP start-up is that compliance with all the obligations of PATS can be an expensive business, but the consumer perception may be that non-PATS compliance indicates a service that is inferior to that of the regular public switched telephone network (PSTN).

By contrast, in the US the VoIP regulatory debate has majored on issues such as its classification or not as a regulation-light information service, states' jurisdiction, taxation and, particularly, access charges. The US debate is on-going. Last November the Federal Communications Commission (FCC) ruled that VoIP was outside the scope of most state regulation, but at least two states are appealing this ruling.

A landmark FCC VoIP vote also has to be taken on or before March 22 on a 'forbearance petition' filed by Level 3 at the end of December 2003. This asked the agency to reaffirm that legacy switched access charges do not apply to VoIP. "In granting this petition, the FCC would foster development and growth of new VoIP applications by reducing the regulatory uncertainty that currently surrounds Voice over IP", Level 3 ceo James Q. Crowe said in January 2004. "VoIP promises to give consumers an unprecedented level of control over their communications services, and to make American businesses even more productive and competitive in the global marketplace. This will happen, however, only if VoIP is allowed to develop without being force-fit into outmoded regulatory constructs".

The fear now voiced in some quarters is that the vote will go against Level 3 and the business case for the VoIP service provider community be threatened by the imposition of access charges.

Not all there yet
You can leave aside the possibility that it might actually be incumbents who take charge of the VoIP bandwagon. You can play down the commercial implications of uncertainties in the still developing VoIP regulatory landscape. Even so it remains that case that VoIP isn't yet a superior (or on some counts even a like-for-like) replacement for circuit switched TDM voice. You might consider that Telstra is only articulating a vested interest when it argues that VoIP is not yet a replacement for its PSTN network and states that the residential broadband IP voice service the company is now testing is expected to be initially attractive as a second or additional fixed line. Telstra is, however, deadly serious about VoIP, putting it at centre stage of its planned A$1.5bn Broadband Multi-Services (BMS) project.

An initial weakness of VoIP was, of course, its quality of service (QoS) when run over the public network, a deficiency subsequently addressed by mechanisms such as multiprotocol label switching (MPLS). VoIP QoS has now improved dramatically, to the point where users often characterise quality as 'really good' or even 'superb'. Whether this actually means 'better than the PSTN', as 'good as the PSTN' or 'not as inferior as we anticipated' is moot. If wireless forms part of a call, the answer to this question can be of some consequence. "There are still voice applications that cannot live with the latency and/or quality degradation of VoIP and so will still need non-compressed access", points out Paul Strawbridge, md of telecommunications engineering company Patapsco Communications. "For example, voice originating through a cellular network already, as we all know, can have quality issues. Subsequently passing this over an IP network using VoIP compounds this".

The service repertoire of VoIP is getting much better, even though people generally don't use, or sometimes even know about, anything like all the features on a modern TDM PBX or even a Class 5 central office. Trouble can start, though, when these taken-for-granted features and services - even down at the level of emergency number access, security and geographic numbering - aren't available. Which, presently, quite some aren't.

Then there's the not inconsiderable issue of legacy and TDM equipment, trillions of dollars worth of which have been installed and continue to function and support bread-and-butter revenue-earning telco services and critical corporate business applications. This installed base is still being added to. According to the Vertical Systems Group market research company even the frame relay market is proving resilient as enterprises migrate towards IP VPNs and Ethernet. And while it may not be sexy, as Strawbridge notes, ISDN is still being deployed in large areas of the world and there is a huge installed base that runs many applications that IP can’t support. "Just because it is not the leading edge doesn’t mean it, or the applications that use it, go away", he contends. "Users and carriers still have to deal with the reality of what’s already installed and not get too carried away with the VoIP hype".

Migration aggravation
Any step-change forklift replacement of existing legacy and TDM systems and equipment with VoIP/IP/MPLS is not envisaged, either from a practical or economic standpoint. Operators and enterprises are accordingly devoting much effort to devising and implementing optimum migration strategies, and companies such as RAD Data Communications and Patapsco Communications make a healthy living out of commercialising appropriate enabling solutions.

As you might anticipate, migration is not as easy to do as it is to say. "The reality is that many organisations have systems and applications that just don’t run over IP at all, or they do not work well with it", states Strawbridge. "Modems and faxes seem to be a particular headache, with migration of these to IP not always going smoothly".

Synchronisation and timing, in which RAD is a leading specialist, is a particular bugbear when attempting to move TDM services over IP. According to Dr Yaakov Stein, chief scientist at RAD Data Communications, so fundamental is the existence of a hierarchy of highly accurate time synchronisation for the proper functioning of TDM networks that today the major carriers no longer employ timing and synchronisation experts. "However, now we're trying to send TDM over non-TDM structure. We're sending TDM over, for instance, MPLS networks. And MPLS networks, unlike TDM networks or even unlike ATM networks, do not send any kind of signal from which you can receive timing", explains Stein. "If we don't actually get good clocking, it blocks any kind of TDM over IP solution".

The unrealistic alternative is to try and go to VoIP in one jump. "If you do that, everyone, over night, has to take all their PBXs and switches and TDM equipment and throw it out and buy softswitches and VoIP end devices - and more expensive telephones perhaps", says Stein.

Which isn't going to happen. In fact some analysts estimate it could take up to fifteen years to change out TDM to VoIP entirely. It could be much sooner, but it won't be tomorrow. Or the day after…
John Williamson