Evolution or Revolution

Internet telephony providers are here. And they're not going to be quiet any more

Somewhere, deep within your networks, a movement is growing. It's not in your old copper wires. It's not in the ground. It's somewhere in a nebulous cloud, floating above you, looking innocent and airy rather than foreboding and waiting to break like a thunderstorm. Oh sure, you can try to scare it off. Tell it it lacks quality. Say it's just cheap so it, too, will pass. But IP telephony-and the people who want to offer it-aren't going away soon. So now you've got to decide: Let them be your friends, or declare war.

The thing is, for the most part, these people want to be friends with you. Not one CEO of an IP telephony service provider will tell you that the public switched network is going to disappear. Instead, they say the two types of networks will eventually integrate. Several of them use the analogy of the personal computer revolution of the early 1980s to show how this integration will happen.

"The PC didn't replace the mainframe computer. It still exists. It's just that the PC was the one to increase in market value," says Kevin O'Hara, chief operations officer of Level 3 Communications. "There are too many circuit switches out there-the [public network] won't disappear. But instead, another network will start to mitigate and dwarf it in value" (Figure 1).

"Certainly the PC was destined to replace the mainframe," says John Edwards, CEO of I-Link, an IP telephony provider. "But it was displaced. It was creatively destroyed. More intelligent networks will creatively destroy the existing ones."

But wait a second before you start cringing at that destroy word-these guys understand where you're coming from.

"[Telcos] have got a legacy to protect," Edwards says. "They've become somewhat of an anchor. But change favors us, and the incumbents hate change."

Of course, be it a revolution that surmounts these small service providers into the forefront of phone technology or a quiet evolution, it hasn't happened yet.

What you're seeing now is the preparation. At telecom conferences, at seminars-it's the IP telephony lectures that are standing room only. Incumbent carriers are quietly testing the technology in their labs, waiting for that open door to long-distance. The small IP telephony-only service providers are partnering with every vendor they can so they can be right there when the explosion takes place. And the vendors are perhaps the most active players of all-trying to create the interoperable gateways to make these networks work and open the technology to a vast new array of IP-based services.

Even analysts can't keep up with changes in gateways over the past year. After being given the task of updating a report on IP gateways that was released last year, Andrew Pierog, a research analyst at Frost & Sullivan, realized that not only the information but also the actual definitions and categories of gateways were already extremely outdated.

"Everyone thought the interoperable gateway would take off, but instead everyone's waiting for the second generation of them," he says. "The growth that was predicted for 1998 will really take off in 1999."

Small service providers are working with vendors to help the next generation develop more rapidly but that "it's kind of hard when you don't know what's going to happen," says Pierog.

He adds that gateways are just the beginning. "Everyone loves their gateways now, but it's just going to be another tool. The real force in this industry will be in the intelligence and the applications."

New applications will arise once quality improves and once access charges and settlement charges make the draw of cheap long-distance calling a lesser selling point for IP telephony, says Karl Duffy, director of telecom services for Killen & Associates.

"Right now, there's a window of opportunity to offer international calls for next to nothing," he says. "But cheap phone calls are not enough. [IP telephony] will also offer the opportunity for other services."

You say you want a revolution? Take a look at a few companies that have a head start.

Not unlike many new industries, small entrepreneurial companies are making the greatest strides in the IP voice market. However, the promise of instant riches is drawing together a menagerie of companies from the Internet service provider and traditional telecom market.

One of the first companies to focus solely on IP telephony, Delta Three handled 17% of IP voice traffic in 1997, according to Frost & Sullivan. Considered to be a pioneer in the industry, it was the first company to introduce phone-to-phone IP calling, says Elie Wurtman, CEO of Delta Three (Figure 2). Because its parent company, RSL Communications, owns a global network, Delta Three is able to terminate voice-over-IP calls all over the world-even where it doesn't own gateways. Its own dedicated IP telephony network handles traffic in 30 countries and provides capabilities for PC-to-PC, PC-to-phone and phone-to-phone calls.

Although the company's core business has been providing capacity to carriers, Delta Three recently entered the consumer arena, Wurtman says. "Our target population is people who are already Internet users-not because you need a PC to use the service, but because we tie a lot of information onto the Web," he says.

This information is value-added services, ranging from voice mail, e-mail and fax mail, to IP conference calling. Delta Three also sells its phone-to-phone software and calling cards on-line.

According to the company, the advantage of working with RSL Communications has been the ability of the two companies to integrate their packet- and circuit-switched technologies. Delta Three also has multiple "global partners," which include on-line content providers, ISPs and vendors. In working with companies such as Ericsson, Delta Three can get very close to the technology, Wurtman says.

"We're looking for reliability, and it doesn't exist in the IP world yet. Our goal was to leverage [Ericsson's] experience to make much more stable products. That's what will take the market to a much higher level," he says.

The company is also working with VocalTec to develop termination services.

Another strong player is IDT, which primarily markets its key product, Net2Phone, as a way to make low-cost long-distance calls. The service is PC-to-phone or phone-to-phone, with the latter becoming more popular, says David Greenblatt, chief operating officer of Net2Phone for IDT. The company is the only one to "mass-market" this service, he claims, targeting business customers and consumers, as well as other telcos. "We're not restricted to any type of customer. It's for anyone who makes a phone call," he says.

Approximately two-thirds of IDT's business is direct sales, while the remainder is in reselling to telcos, Greenblatt adds. The company offers extensive customer service over its IP backbone, which serves 62 cities in the U.S. and 12 in Korea.

What makes IDT unique, according to Greenblatt, is that it has created a gatekeeper in-house. "It runs our network around the world," he says. "It can manage hundreds of these kinds of calls-and it's making us a leader today."

The company has also ventured into voice-enabled e-commerce with its Click2Talk, which allows businesses to set up buttons on company Web sites that offer customers a chance to speak to customer service or sales departments. International customers can do this through an 800 number, which should also mean lower rates.

Several smaller "next generation" telcos also advertise their services as a way to call all over the world at little cost (Figure 3). One way they can differentiate themselves is by offering value-added services, says Bob Bryson, vice president of product marketing for I-Link. "We believe arbitrage and the access charge issue is short-lived," he says. "We need other services that fit our customers' business needs."

The company focuses on the small office/home office and small-business market, and offers a unified messaging capability on its platform that includes voice mail, teleconferencing and one-number dialing.

"The basic idea is to take existing services and provide them within the IP environment," I-Link's Edwards says. "[IP telephony] has gone from very exciting for all the wrong reasons to very compelling for all the right reasons."

For now, however, many operators are content to go after the consumer market by selling cheap long-distance minutes. In fact, many small operators have put off chasing corporate customers until the quality improves and users don't have to dial up to 25 digits to complete a call.

"One of our biggest problems is the number of digits you have to dial," says Doug Hanson, president of Rocky Mountain Internet, a Denver-based ISP offering IP voice. The company makes a point of informing consumers that the quality of their IP calls will not be at the same level as "regular" phone calls. "Businesses in general don't want that inconvenience. If you can neutralize that for them, they'll use it, and frankly that hasn't happened yet."

Other companies such as Internet Global Services act like middlemen, reselling their services to other ISPs. "We understand that many ISPs cannot afford to build a network like this-and that makes the issue of voice over IP seem impossible to them. That's why we offer our service to them," says Michael Gorton, CEO of Internet Global Services.

ISPhone, a Traverse City, Mich.-based ISP, is taking the concept even further down market by targeting ISPs with no more than 20,000 customers. Under the company's model, ISPs pay a fee to access ISPhone's network of Cisco routers and also are given exclusive rights to terminate traffic in their local areas off the network.

"Right now, we're concentrating on national minutes," says Kenneth Moody, vice president of ISP operations for ISPhone. As of late October, 30 ISPs had signed up for the service, which was expected to launch this month.

At the other end of the size spectrum, large ISPs such as PSINet are getting into the game. PSINet which will roll out IP voice services in three stages, the first of which-a PBX-to-PBX service designed to replace private circuits-was launched in October.

"PSINet owns everything above Layer 2, and because of that we can do some unique things with the service," says Tim Suh, engineering consultant for PSINet. "We can do truly private networks."

More importantly, by owning the backbone portion of the network, the company can control quality. "We don't have packet loss or latency [service level agreements], but because you're avoiding the congestion on the Internet, you avoid all the latency and dropped packet issues," says Suh. "Because we own the backbone, we think we have an advantage."

A few companies, such as Level 3 and Qwest Communications, have built their own private networks to handle voice over IP. Level 3, which is building a nationwide IP-based network to be completed by 2001, presently offers only IP telephony call termination for ISPs within its network, but it plans to provide call origination in 1999, O'Hara says.

Level 3 has developed its own gatekeeper, which it calls a soft switch, because it makes circuit switches unnecessary. The soft switch connects to the public network and directs traffic to databases when needed. "It allows us to originate and terminate calls with no change in behavior on the customers' part," O'Hara says. The soft switch has been certified by Bell Atlantic for three markets in its territory, he adds.

The company has also worked to develop protocols with which to use the switch. "We initiated a technical advisory council, included all the leading vendors and ended up with a set of protocols that will allow IP to move into the mainstream without having to rely on proprietary devices," he says.

Like many other IP telephony providers, Level 3 works very closely with telcos to establish their IP telephony business. O'Hara says the company tries to be very flexible with regional Bell operating companies. "Our going-in proposition is that we're not going to ask anyone to change how they do business," he says.

Not surprisingly, traditional telcos don't see an uprising of IP telephony providers supplanting their networks. Or, as Kathleen Early, vice president of Internet services for AT&T, says, "This is going to be an evolution, not an event."

Of course, AT&T is the one large telco whose thoughts on the subject are now widely known. In mid-October, the interexchange carrier introduced a Global Clearinghouse that will help ISPs and PTTs establish IP voice services (Telephony, Oct. 12, page 6). It also introduced Web, consumer and business strategies for providing the service.

Analysts were amazed to hear that such an established telco would step so far into the IP voice realm. The move shows AT&T is one of the more aggressive telcos, says Killen & Associates' Duffy. It also could pose a threat to the small IP voice providers, he adds.

"One of the most challenging aspects of being an [Internet telephony service provider] is being handled by AT&T," he says.

It will also allow the carriers and ISPs that work with AT&T to circumvent settlement charges in other countries because of the IXC's strong relationships with PTTs there, he explains.

AT&T will concentrate on deploying phone-to-phone service. "Right now, we feel that both the PC-to-phone and PC-to-PC services have quality issues that limit their appeal," she says.

The phone-to-phone service that AT&T introduced in October uses multistage dialing and debit cards. Most of the calls will go across a private IP network, allowing the company to control quality.

But again, AT&T is the exception. Other IXCs, which have the most to lose in the early market of IP telephony, have approached the market more gingerly. MCI WorldCom executives say they're exploring the IP telephony market, but publicly the company continues to pursue its On-Net program, which provides corporate users a single network to carry voice, data and video. In many cases that network is based on frame relay, which can carry IP voice. As a separate product, MCI WorldCom is still some time away from announcing its IP voice intentions.

"Voice over anything including IP is a peripheral game right now," says Robert Smith, a senior marketing manager for MCI.

Although not completely comfortable with the revolutionaries, RBOCs at some point may find their biggest allies in the new carriers, for within IP telephony is a relatively fast and easy way to offer long-distance services. Early on in this love-hate relationship, the focus has been on access charges and reciprocal compensation.

Both BellSouth and U S West announced in September that they would impose fees on ISPs for connecting phone-to-phone IP calls from local customers. Ameritech also said it would refuse to pay reciprocal compensation to CLECs that link ISPs with its local network. At the same time, Bell Atlantic said it would install a gateway in New York to supply local connections for international IP phone calls terminating there.

Fred Chang, vice president of data technology and strategy at SBC Communications, says the carrier will initiate big plans for voice over IP-just as soon as it gets a full-blown data network in place. By the end of 1999, SBC will have spent $6.5 billion on data enhancements over five years. Chang says that deploying IP telephony will be consistent with its data service offerings.

"The importance of IP to our data strategy equals the importance of our data strategy to our overall outlook," he says.

By the end of 1998, the carrier will offer an IP/VPN service, and it plans to roll out other voice applications such as push-to-talk over the next year, Chang says.

Other carriers have remained quiet. That's not surprising since RBOCs can't offer long-distance, which has been so integral to selling voice over IP. Ameritech is working on the technology with European carriers TeleDanmark and Belgacom, says Andy Schmidt, a product manager in data services. The fundamental problem with IP telephony is providing adequate performance, he says.

"The performance issue is always going to be important," he says, naming issues such as latency, jitter, packet loss and layer interface. "We're almost there, but there's no such thing as a complete solution yet."

Ameritech has conducted field tests at universities and large corporations, Schmidt says.

Even the IXC forerunner admits having difficulties with the service. "Voice is the toughest application to move to IP," said AT&T's Early at Internet World in October. "It's unforgiving."

Stepping into that breach are more than two dozen IP telephony gateway vendors, all claiming to have solved the technology's toughest issues. A close look at the technology, however, reveals a category of products that has evolved so rapidly that even putting a definition on gateways becomes difficult.

Pierog says that when Frost & Sullivan released a report on the gateways less than a year ago, the distinction made between the products was "stand-alone box" or "integrated." Already, these categories no longer apply, he says.

"Stand-alone still means the same thing, but integrated can have several different definitions," he says. "Integrated can mean integrated into the present network, or it can mean integrating a gateway into equipment. It can also mean integrating technology into the gateway."

His solution to this categorization was to define equipment by number of ports. New players in the market further complicate research on the subject-Pierog says a year ago there were 23 voice-over-IP vendors; now there are 39.

In addition, vendors are coming at IP telephony from different directions. Companies that have long worked to develop phone technology will differ from vendors creating products for the data world.

"Lucent comes from a telco background, so they'll approach IP telephony from a telco-solutions perspective, whereas Cisco will approach it from a data background. That's why they're putting gateway technology into their routers," Pierog says.

Cisco Systems wasn't even on the playing field a year ago (Figure 4). That doesn't mean the company wasn't working on the technology, says Mark Bakies, product manager in the voice-over-IP group at Cisco.

"We weren't as clueless as people thought. We were trying to figure out how to embed gateways into routers," he says, adding that the company worked on its offerings for two and a half years before they were introduced.

Researchers at Cisco had watched several types of data technologies become voice-enabled, and they didn't want to release anything until they could be sure of its quality, Bakies says.

"We spent a fair amount of time trying to solve a critical problem. We decided that if we could not build a packet-switched network that could deliver [public network] quality, then it wouldn't fly," he says.

The company now has products that are well-positioned to integrate with other voice technology, Bakies adds.

Lucent's idea of integration is to create a platform with which other vendors can operate a gateway, says Kalpana Sheth, director of marketing for Lucent's elemedia venture. However, the manufacturer also makes gateways and gatekeepers with applicable software, she says.

"Our goal is to have a product that can expand with levels of use," she says. She explains that the first level of IP telephony technology is to provide cheap service, the second level is to make it simple or "transparent" to the user, and the third level will be to add multimedia functionality.

Lucent, like several other hardware-focused manufacturers, has partnered with vendors such as VocalTec that lean toward the software side, to work on the next generation of IP voice products. Such cooperation should boost the speed of interoperable products and increase quality, Pierog says.

Meanwhile, IP voice guru Jeff Pulver predicts in his "Pulver Points" that 1999 may be the year of gateway interoperability. "Recent announcements such as VocalTec/Lucent and Motorola/Inter-Tel are a good sign, but we're still waiting for actual interoperable products," he writes. "Imagine if two carriers got together and there was close to zero interoperability between their switches?"

Other recent announcements include Siemens Information and Communications Networks and GRIC Communications Inc., which will work together to combine Siemens' Hicom Xpress Telephony Internet Server with GRIC's line of client/server software.

Smaller vendors, however, seem to be focusing more on new applications for voice over IP.

In a seminar at the Carrier IP Telephony ComForum in October, Dominque Linden, product marketing manager for Dialogic, compared IP telephony to a space shuttle, with cheap phone calls serving only as the boosters. "They help, but they won't help land it safely," she says.

Already, voice-over-IP gateways can provide call center integration, Internet call waiting, fax over IP and unified Internet messaging. Voice over IP will result in the creation of IP-based call centers that will start with push-to-talk technology and result in an "open IP world" where voice over IP, a PBX, fax and other services will all be provided by one server and one service provider, Linden says.

Call center functionality is just one of many technologies that carriers are asking for, says Robert Kennedy, president and chief operations officer of IP-voice vendor NetSpeak Corp.

"The first implementations were to save money, but now a great amount of attention to applications is turning this into a serious business," he says.

An example of a new such driver, he says, is "Click'nConnect," a service MCI WorldCom has developed with NetSpeak's technology. The service includes the ability for "least-latency routing," or routing to the closest gateway, Kennedy explains.

The vendor is funded by companies such as Bay Networks (now part of Nortel Networks) and Motorola, which has invested $90 million-the largest investment by a partner in this industry today, he says.

The key for developing equipment in this market is creating products that can change with the industry, Kennedy says. "Seventy-five percent of the cost for a carrier is maintaining the network, so they need to buy equipment that will be in for the long haul. And that is clearly packet-switched."

All of these vendors, like the next generation telcos, see an integrated network of the future.

"By the very fact that AT&T, MCI and Sprint are building out packet-switched networks, it's clear these are the networks of the future," Kennedy says.

The key to realizing this network, many people say, is recognizing that IP telephony is more than just a way to offer cheap long-distance. "There are so many more advantages," says IDT's Greenblatt. "There are multimedia capabilities-capabilities for manipulating data at the same time. There's the fact that you'll only need one network. And there's the ability for service providers to use their resources better."

"IP voice is more efficient than the [public network]," adds Gorton. "In the future, all voice will be packet-based telephony. It just makes sense. Hundreds of people being able to use the same piece of copper saves money."

It could take up to 10 years, but carriers will move to this type of network, Strauss says. "It will take over, for one simple reason," he says. "It takes one-tenth of the circuit of a simple growth voice channel. It's much more practical. It offers great convenience, and we have the capability to make our technology go that way."

For Wurtman, an integrated, hybrid network is a simple fact of the future of telephony. "We see ourselves as the modern phone company," he concludes.

Do note, however, that many next generation telcos would prefer a period of peaceful coexistence between IP and the public network because, most admit, a flash-cut to an all-IP network would be a disaster. "IP couldn't handle the calls today even if the technology were there," says Hanson of Rocky Mountain Internet. "The quality issue alone would prevent it from taking over too fast."

Pierog agrees that coexistence would be ideal. "Maybe it will just help the data and the voice industries blur, and then we'd have the best of both worlds," he says.

So should you give peace a chance?

"Well I guess that's a nice idea today," he adds. "We'll see if that happens. "

Growth in IP telephony traffic is driven by the following factors: * Possibility for service providers to practice price discrimination

* Toll bypass

* Possibility for ISPs to practice usage-based pricing

* Creation of IP telephony settlement firms

* Need for enhanced call centers

Growth in traffic is restrained by: * Lack of interoperable equipment

* Lack of true carrier-grade equipment

* Nonstandard accounting and billing capabilities

* Regulatory uncertainties

The IP telephony service market is affected by the following trends: * IP telephony services are mostly offered as prepaid

* Most service providers have their own managed networks

* Settlement firms are expanding and peering

* An increasing number of Web sites incorporate click-and-talk buttons

* Service providers are introducing Internet call waiting and IP telephony VPNs

Source: Frost & Sullivan, January 1998