The Road to VoIP

FRAMINGHAM (03/06/2000) - Voice-over-IP vendors are flooding the market with IP phones, gateways, converged appliances, stand-alone boxes, expansion cards, expandable routers and IP PBXs. Carriers are racing to convert their networks to IP so they can offer voice-over-IP services.

But how's this for a reality check: Less than 4 percent of the 70 billion voice minutes logged last year were over IP, according to Probe Research.

And for good reason. Although voice over IP offers potential cost savings and opens up the possibility of new applications, as things stand today, there are significant barriers to voice-over-IP implementation and very few real incentives.

Let's start with cost savings. Voice over IP is supposed to save you money by eliminating expensive long-distance calls, often referred to as long-haul toll bypass. But toll charges have dropped dramatically in the last two years. And an Internet telephony system in many cases requires upgrading to a switched LAN, which adds a significant cost.

Bottom line: You're simply not going to save a lot on long-distance charges by switching to voice over IP.

You could try running IP voice over an existing WAN that has excess bandwidth.

This strategy works within a closed, switched network, a WAN, LAN or virtual private network. In closed systems, you control the utilization and quality of service (QoS). Of course, that raises the issue of why you're paying for excess bandwidth in the first place. Maybe you'd save more money reducing your WAN bandwidth charges.

Barrier No. 1: Management tools for converged networks won't be available for two to three yearsThe more enticing potential cost savings from voice over IP is the opportunity to slash network management costs. Put simply: Most enterprises typically have separate phone and data networks with completely different wire infrastructures, separate equipment, operations centers, trained administrators and distributed end-user devices. Voice over IP gives you the ability to consolidate them into one network.

Specifically, with Web-based management of voice switches and PBXs, the use of directories to handle adds, moves and drops for both the phone system and network users, and Ethernet phones that plug right in to 10Base-T jacks, equipment is consolidated, administration is streamlined and overall costs are reduced.

Voice-over-IP proponents point out that 70 percent of ongoing network costs are related to administration, and companies that are consolidating phone and data networks could slash their network administration costs by more than half.

But those estimates may be very optimistic, according to Gartner Group Inc. analyst Kathleen Simpson, who points out that voice and data are historically different mindsets and skill sets.

She says management tools with common interface, alarms and maintenance capabilities to handle a converged network are two to three years away. Without them, it's difficult to realize lower management costs.

Barrier No. 2: Applications are six to 12 months awayAside from cost, new features are a second motivation for implementing voice over IP. For example, you can integrate with voice gateways to create Web applications that are logic-driven, although these applications generally need to be custom-developed.

E-commerce customers could browse to a certain point and then connect by phone to a sales represent-ative who sees the browse trail and customer history.

Also, virtual phone systems can be created providing voice mail, speed dial, conference calling and other typical phone features, no matter where you are in the world.

If cost isn't a real motivator for implementation, these new features can be.

International Data Corp. Analyst Paul Strauss says the benefits of an IP-based phone network include shared conferencing that doesn't require operator setup, simultaneous voice and Web presentations, combined inboxes and aggregated billing. He says larger businesses need call centers with automatic call distributors that forward calls across states and allow incoming calls to be routed to remote centers or to homes.

He sees these new applications as the drivers for voice over IP and says we can expect to see such products coming to market late this year or early next year.

Barrier No. 3: The Internet is still too unreliableThe general Internet infrastructure isn't stable enough today to support voice calls. Because the Net without QoS is unpredictable, the resulting latency produces poor or unacceptable voice quality.

You need QoS network protocols such as Multi-protocol Label Switching (MPLS) and IP type of service (ToS) to sort voice from data packets and guarantee that voice packets are received in an acceptable amount of time.

The bottom line: Companies need a packet-switched network with QoS protocols for voice with IP telephony applications to reliably match a public switched telephone network (PSTN).

Barrier No. 4: Standards are still evolvingAccording to analysts and vendors alike, a significant barrier to ubiquitous IP telephony is the elementary state of the interoperability standards in the voice-over-IP arena. The reason is that while the standard is there, the interpretation is still loose, which is typical for evolving standards.

The International Telecommunication Union has released Version 2 of the H.323 interoperability standard, an upgrade to the original H.323. New features included call hold, call park and pickup, call waiting, message waiting, and some fax and multimedia broadcasting capability. These are features that basically map voice calls over IP and standardize call connections, allowing calls from different systems to interoperate.

However, H.323 Version 2 provides only the foundation of functionality for call interoperability. What's lacking is a common standard for the rich set of potential IP telephony features and applications. The bottom line: Getting voice conferencing on your system to work with your business partner's on another system will be difficult.

However, this situation could change with the new Media Gateway Control Protocol (MGCP) standard. MGCP has been submitted to the Internet Engineering Task Force and is backed by vendors and carriers. It provides some level of interoperability. Calls that interact with applications such as Web pages will be able to interoperate with other IP calls. For example, a user of a Cisco phone network could set up a voice conference with users of Nortel Networks and Lucent phone networks.

Additionally, standards for QoS are critical for successful voice over IP. IP ToS, part of the IP packet header, can assign special treatment, such as routing, to packets. ToS signaling, Real-time Transport Protocol header compression, and MPLS are standards that work to reduce IP overhead by directly reducing bandwidth requirements. That could indirectly reduce latency and ensure QoS to the level required for voice calls.

Despite these concerns and others, such as security, momentum is building.

Carriers have generally accepted that circuit-switched networks will be replaced with packet-switched IP and are gradually implementing and upgrading infrastructure to handle it. This movement in the long run will eliminate problems such as poor QoS, the need for conversion, and mapping and translation devices. ISPs likewise will be providing IP dial tones and call management services.

If you have a working phone system now, you may want to wait. Full IP telephony outside of a closed system is still a way off, but it's coming.

On the other hand, if you have the need for IP telephony features now or think you could realize some benefits from implementing a voice-over-IP gateway, sit down and work out your own convergence strategy. Consider the benefits from toll bypass and identify which gateways are available to meet your needs.

You could consider using an IP PBX internally and switching to PSTN for outside communication. Or decide on a collection of individual or hybrid products that will handle your current needs while positioning you for full IP telephony in the future. Designing your converged voice and data strategy is not a simple task, but there is no shortage of available options.

Peter Clegg is a technology writer and consultant associated with Key Labs and Beacon Strategies in Utah.

Join the newsletter!

Error: Please check your email address.

More about GartnerGartnerGatewayInternational Telecommunication UnionInternet Engineering Task ForceLucentNortel NetworksProbe Research

Show Comments