October 1999

Challenges Facing Internet Telephony Service Providers

BY SEAN PARHAM

The hundreds of Internet Telephony Service Providers (ITSPs) that have emerged during the past two years face a daunting challenge: Keep stealing minutes from traditional telcos, continue to mine ethnic/foreign markets, maintain consistently high voice quality, optimize network infrastructures, and introduce new service offerings. Overall, the challenge is to grow or die.

The recent advent of voice communications over IP networks (VoIP) shook the telecommunications industry to its very roots. Once companies began using their existing corporate intranets to make long-distance voice and fax calls, some startling new numbers emerged, e.g., what used to cost $3 suddenly cost 30 cents. Rarely do industries see such radical changes in their economic cost models.

These companies, who had been customers of AT&T, Sprint, MCI WorldCom and the other long-haul giants, suddenly realized that they could achieve a tenfold or better reduction in long-distance costs over their own networks without adding infrastructure. This represented an incredible windfall, and an equally fearsome threat to the traditional long-haul carriers.

The cost reductions associated with VoIP are easily quantifiable. Motorola recently conducted a study comparing the costs of certain file types being transferred over a regular voice network versus an average internal IP data network. The results indicated that an average 10-minute international phone call may be transmitted over a voice network for $4, or 40 cents per minute. The same phone call over the private IP network costs 14 cents or 1.4 cents per minute.

EARLY VoIP MARKET DYNAMICS
There were plenty of profits for the first ITSPs to arrive on this untapped scene. Buoyed by their prospects, they made seemingly “too-good-to-be-true” propositions to large and small companies, companies that were initially skeptical, but frequently won over by the irresistible savings. With $60 billion in VoIP revenues forecasted by 2005, ITSPs began to sprout up all over the place.

These hungry young companies immediately turned their attention to international markets such as Eastern Europe, Asia, and Latin America, where long-distance voice rates typically ran up to 85 cents per minute. By offering rates in the range of 35 cents per minute, and targeting concentrated ethnic communities in U.S. cities, ITSPs were able to offer reduced rates and still turn a handsome profit.

As is usually the case, however, the too-good-to-be-true scenario turned out to be just that. Many early VoIP services from the new crop of ITSPs were less than successful because voice quality suffered as transmissions were passed from one unreliable IP network to the next. Not even big savings were enough to make corporate customers compromise on the quality of their voice networks. Good voice quality is like a birthright — and very few are willing to renounce it.

This negative scenario has now largely reversed itself, as many ITSPs have developed their own private, managed IP networks, which means they can offer reliable, predictable quality of service (QoS) for end-to-end VoIP calls. They are maximizing the value of these networks by employing data compression techniques that allow them to condense 10 calls on the amount of bandwidth that used to be required for just one voice call.

CURRENT CHALLENGES
Still, problems and challenges remain. While network performance improves and the price of sending bits through a data network decreases, quality suffers on calls not transmitted through private networks. Even now, when using a general-purpose, or multi-use IP network, organizations have little to no control establishing QoS and priority standards for their data and multimedia packets.

On the positive side, the concept of the “active network,” wherein the network quickly responds to the content of a packet, rather than passively routing or switching packets, is becoming a reality in some respects. However, as VoIP moves inexorably closer to becoming the de facto standard for voice transmission, and as service providers offer not only VoIP, but video conferencing and other multimedia services over their IP networks, true quality, prioritization, and intelligent routing need to be applied to all levels of IP networks.

THE POTENTIAL OF PERSONAL NETWORKS
Among near-term possibilities for new and enhanced services, personal networks stand out. Personal networks will allow users to communicate in a variety of formats, such as voice and text, from a variety of locations — their cars, offices, or homes — via a variety of devices, including PCs, laptops, cell phones, and pagers.

In addition to offering personal networks, ITSPs will also be associated with other IP-based phenomena such as Web-based call centers, software PBXs, and voice, data, and application conferencing.

LIVING ON THE EDGE
In order for ITSPs to realize their full potential, major network changes must be planned, subsidized, and approved. Removing the intelligent QoS function from the network core and moving it to the edge of the network, where service providers interact with customers, is the key change on the horizon. This makes sense because the edge of the network is meant to provide revenue-generating services. The customer perception of QoS largely depends on the performance of the edge. Moving intelligent QoS there not only gives organizations more control of their IP infrastructures, but it also reduces their costs of responding to the demands of fast multimedia services.

Traditional IP routers and switches were designed to switch packets rapidly through data networks. ITSPs require the addition of quality, applications, and multimedia services. Although current IP routers are capable of transporting voice, video, and data packets efficiently over carefully engineered networks, they lack the flexibility to deal with the dynamic nature of fast-moving multimedia services. This is changing with the arrival of the first generation of intelligent multimedia devices that coordinate and optimize mixed audio, video, and data traffic across IP networks. As successive and less expensive generations of these sophisticated products appear, IP networks will consolidate their hold on all forms of voice and data traffic.

DOING BATTLE WITH THE BEHEMOTHS
ITSPs will need every possible technological advantage in order to survive the VoIP onslaught of the incumbent local, long-distance, and international carriers. These behemoths have quickly moved from being threatened, to taking a variety of actions. In the regulatory arena, some of them have filed petitions and proposed that Internet access calls be billed by the minute. On the business side, they are also actively seeking vendors who can help them enter the VoIP arena. Despite these actions, however, they face a diabolical dilemma: Do they undermine their core voice telephony business by offering VoIP, and if so, to what degree?

The big telecom companies may be late to join the fray, but Microsoft was also late when it got serious about the Internet. Despite that handicap, however, the company quickly carved out a huge market share by applying its vast resources to the market. AT&T, Sprint, et. al. have similar resources and a lot of business acumen at their disposal. ITSPs will need well-formulated business plans and a solid installed base of customers as they battle the big telecom companies.

When they first hit the market, ITSPs were fortunate to be presented with abundant opportunities that are rarely available. Now that the advantages of early entry have given way to the exigencies of intense competition and the demands of cutting-edge technology, they must move forward with advanced services faster than ever in order to succeed.

Sean Parham is Director, Inter-net Products Operation (IPO), for Motorola’s Internet and Network-ing Group. IPO is working on a next-generation product that offers a common platform for delivering multimedia and QoS into wireless, cable, networking, and service provider markets. For more information, visit the Internet and Networking Group Web site.

BY SKIP CARLSON

Voice over digital subscriber line (VoDSL) presents a tremendous opportunity for service providers in the converging data and voice communications market, including competitive local exchange carriers (CLECs) and incumbent LECs (ILECs). Its roots trace back to the Telecommunications Act of 1996, which, bolstered by subsequent enhancements, has and continues to effectively increase competition in the telecommunications industry.

Since 1996, high-speed Internet access has been the primary market for new service providers because of the ever-growing demand from both business and residential customers. In recent years, DSL has emerged as the most affordable method for carriers to serve both types of subscribers.

For businesses, symmetrical DSL (SDSL) is the preferred method for high bandwidth access, replacing T1 lines that can be four- to five-times more expensive. But for voice, the application of choice thus far has been asynchronous DSL (ADSL), primarily because it offers single-line POTS support. While well suited for single phone line and residential applications, ADSL does not fare well in support of the business model. To successfully serve the booming market for small and medium-sized businesses, competitive services have to include multi-line voice and data offerings. To achieve this, service providers can now offer voice and data access via a single copper wire using DSL as the physical access transport method.

BREAKING DOWN DSL
Within the transport, a number of technologies are supported to transfer both voice and data. While data transport has been optimized, the simultaneous transfer of toll quality voice needs to be addressed. Emerging VoDSL product solutions do not change the model for data and voice access into the voice network. Instead, they change how data and voice access the network through emerging technology.

The methods used in support of a voice service over DSL vary, as do the vendors who support such service offerings. When looking at this emerging technology, we can learn from past lessons. Specifically, initial voice-over-IP (VoIP) initiatives were not successful because the quality of service (QoS) could not be maintained throughout the network. Also, H.323 in support of VoIP supported voice, video, and data, but maintained too much overhead for simply supporting voice. Data is better off being left out of the implementation and supported in its native format. Video can follow in the data path as long as delay characteristics are kept at a minimum, and QoS is maintained. The compression methods in support of video today are very advanced and optimized for current data standards.

For current VoDSL implementations, several proponents are enjoying success by limiting the scope of the network that needs to maintain the QoS (voice compression and other advancements are enabling QoS throughout the network). One example is that by optimizing the standard VoIP structure within H.323 for voice only, maintaining IP frames, and using Q.931 signaling, the IP model works quite efficiently. This model also enables the gateway to be located anywhere on the IP network and the passage of voice from the subscriber to the gateway occurs over an IP-enabled DSL topology.

This modified form of VoIP is just one method of implementing a VoDSL solution. Another sample approach is to utilize ATM or frame relay virtual circuits, whereby AAL2 is used for creation of voice circuits using Q.931 signaling (a derivative of H.323) in a Point-to-Point topology that terminates in a gateway local to each central office or each Class 5 switch.

Furthermore, some applications support ATM, while also offering traditional signaling methods to interface the customer premise equipment into the central office telephone switch (known as robbed bit signaling). This is something the exchange carriers have used for years, but unfortunately creates a lot of unnecessary overhead on the VoDSL network.

In deploying VoDSL applications, DSL equipment takes on added importance for service providers. A new category of integrated solutions is emerging that supports all VoDSL approaches, including the modified VoIP and the ATM/frame relay-based methods. New single-box DSL Access Routers (DSLARs) are flexible enough to support any type of network architecture where DSL is required. Multiple traffic prioritization schemes can co-exist throughout the platform, and the DSLAR can make decisions simultaneously at both layer 2 and layer 3. QoS is represented through IP layer prioritization, multiple layer 2 priority queues for frame relay or ATM, and standards-based approaches such as AAL2. All signaling occurring between the customer’s voice network and the voice gateway is transparent to the DSLAR, since it is a true communications stream between the customer premise equipment and the voice gateway.

Like Internet access several years ago, VoDSL today represents a significant market opportunity for innovative service providers. However, because there are various ways the technology can be implemented, carriers need to pay close attention to their DSL equipment to minimize interoperability issues and to ensure toll quality.

Skip Carlson is director of product management for Interspeed. Interspeed is a wholly owned subsidiary of Brooktrout Technology, Inc. The company develops single system remote access solutions for local exchange carriers and ISPs who are striving to deliver greater speeds at lower costs. Interspeed’s solutions are specifically designed as robust, efficient data-oriented platforms, inherently better-suited for broadband data communications. For more information, visit Interspeed’s Web site at www.interspeed.com.