IP Telephony: Do the Numbers Add Up? - Technology Information

Telecommunications, June, 2000 by Tracy Venters

The bottom line is revenue. No matter how efficient a network is, it's ultimately services that make operating the network profitable. Is IP telephony the answer?

Since the first softphone hit the market in 1995, IP telephony has quickly evolved from a capability for offering free long distance to Internet hobbyists to a technology now being embraced as a cornerstone of next-gen networks. Initially, the excitement around IP telephony centered on the cost savings realized from transporting voice over packet networks. However, the real economic opportunities lie in its ability to deliver real-time, person-to-person multimedia communications.

One key advantage of IP telephony is the bandwidth savings resulting from the use of packet switching instead of circuit switching. By using algorithms that detect and suppress silence, as well as aggressive compression techniques, a call can be delivered at rates as low as 8 kbps--a significant bandwidth savings over the 64 kbps required for a circuit-switched connection. However, standard IP-based network protocols do not guarantee quality-of-service (QoS) characteristics.

The first defense against poor QoS is often to overprovision the network so there is always plenty of bandwidth, but this results in wasted capacity. A more strategic solution is to deploy networks that implement QoS protocols and methodologies in applications and routers. With either tactic, ensuring QoS raises the cost of deploying IP telephony.

The cost of networking equipment must also be factored in. Traditional circuit switches are massive machines that can be very expensive. By contrast, data-switching equipment is based on off-the-shelf, general purpose computing components that can be an order of magnitude less expensive than a proprietary counterpart. These solutions are not necessarily carrier class, however, manufacturers of traditional data equipment are building in the redundancy and fault tolerance required by service providers, albeit at a price.

One of the biggest arguments in favor of a converged voice and data network is the capital and operational savings resulting from deploying one multiple purpose network as opposed to multiple, single-purpose networks. Integrated communication carriers today often have a voice network and up to three separate data networks for providing IP, ATM and frame relay services. The first costs incurred in building these networks are capital costs. These expenses include the switches and routers as well as the costs involved in constructing facilities to house this equipment. However, the biggest cost is the ongoing expense of network operations. Networks must be engineered and deployed, monitored, repaired and periodically upgraded, and associated personnel must be hired and trained.

It doesn't take an accountant to figure Out the tremendous opportunity for reducing expenses by deploying one network instead of three or more. Although it is likely that collapsing the networks will require increasing the capacity of the multipurpose network, this is still more cost effective than operating multiple networks. Economies of scale are an important factor in networking equipment with prices dropping as port density increases.

The Cost of Doing Nothing

The hype around the economics of IP telephony deserves a closer look. While transport is certainly cheaper, factoring in added bandwidth, the overhead involved in providing acceptable QoS, and the expense of carrier-grade equipment adds to the cost of the traditional data network. Although the aggregate cost of providing IP telephony is still less expensive than the traditional, nonconverged network, the cost of not investing in the new network must be factored in to grasp the real value.

The last decade has seen phenomer growth in Internet usage. A market study from Ovum reported that the resulting ubiquity of IP has made it the de facto global standard for data networking. Ovum predicts that worldwide revenue from next-gen services will grow from $74 million in 2000 to almost $40 billion in 2006 (see Figure 1). According to the report, the buildout will include a steady growth of access services throughout the forecast period from 2000 to 2006. Migration of enterprise networks from other data services to IP virtual private networks (VPNs) will begin this year and accelerate rapidly around 2002 in developed markets. The development of next-gen infrastructure will accelerate the growth of application services around the middle of the forecast period, as the lower investment and risk involved in service development encourages increasing innovation and customization. Applications will become the largest segment of next-gen service revenue around 2005, but will occur earlier in the United States , where deployment of server-based services and media gateways will be particularly aggressive.

The surge in Internet use will come at the expense of existing transmission services. Another report, co-authored by the International Engineering Consortium and the University of Southern California Center for Telecommunications Management, states that data transmission will keep the lead in percentage of usage as more home PCs gain access to the Internet, propelled by e-mail and Web applications. However, telephony services will move to the Internet as well.