Helping copper GROW

Despite some aggressive rollout plans, digital subscriber line has not yet found its sea legs. Will its promise be realized in '99, or is it in danger of an ISDN-like demise?

Digital subscriber line technology represents the promise of a new life for copper. As one of the well-publicized solutions for the well-publicized last-mile problem, DSL must live up to great expectations. It has scarcely arrived on the scene, and already it is enduring heavy scrutiny and doubt.

As often happens in an industry known for overpromising, DSL deployment to date isn't going as smoothly as carriers originally hoped. Stymied by standards battles, DSL got off to a slow start. As that settles down, carriers have to discover how and to whom to market an anomalous service. In addition, competing technologies such as cable modems are giving DSL a run for its money.

Still, service providers are making inroads with DSL. Advocates point out that cable service doesn't top most consumers' lists for product reliability. The market, they say, is simply waiting for an alternative. Now that DSL is slowly emerging, customers might be able to get the bandwidth boost they need.

Dissecting DSL DSL is designed to open up the bottleneck in most networks-the copper link between the service provider and the end user. Because it runs over existing phone wires, DSL is very attractive for several reasons. For example, carriers don't have to run new cable to the home or within the home to deploy service. However, users have to install a DSL modem that talks to a DSL access multiplexer (DSLAM) in the central office (Figure 1). In addition, DSL is an access technology, so it is protocol-agnostic.

Tapping into the unused spectrum, DSL service operates in frequencies outside those used for voice. Voice or analog transmission typically falls in the 0 to 3400 Hz range and transmits at speeds from 9.6 to 56 kb/s. By extending the top boundary, DSL can operate on the same line without interfering with the analog signal.

But to support DSL, the copper wire must meet certain standards. For example, the line must be free of load coils, which are often added to long local loops to improve voice quality. DSL doesn't travel through remote terminals such as digital loop carriers (unless you are using ISDN DSL, which is explained later). Links typically have to be 18,000 feet or less, and the longer the loop, the slower the DSL rate.

The increase in bits moving around the network gave rise to DSL technology. Several types have emerged, each with different speeds and each suited for different applications (Table 1).

Two of the most common are high bit-rate DSL and asymmetrical DSL. Other DSL technologies include symmetrical DSL, IDSL, high bit-rate DSL, very high bit-rate DSL and rate-adaptive DSL.

IDSL was designed to support ISDN, but unlike most DSL technologies, IDSL doesn't run on the same phone line as POTS service, says Frank Weiner, vice president of Paradyne's DSL products group. The ISDN signal is sent through a digital line carrier and connects to an ISDN card in a DSLAM at the CO.

HDSL offers duplex transmission of 1.5 Mb/s. HDSL enjoys a wider deployment than ADSL, with some 2 million HDSL circuits available, says John Griffin, vice president and general manager of the loop transport division at ADC Telecommunications. HDSL is used as a means of delivering T-1 based service over copper pairs. Its symmetrical capabilities make it well-suited for business-to-business transmission (Figure 2).

The HDSL industry has only a few key players, Griffin adds, noting that "we're all in the fifth or sixth generation of the product."

While HDSL enjoys a fair presence in the background, ADSL has generated increased interest. After a slow start, ADSL is popping up on carrier radar screens all over the globe.

"Primarily the winner in the mass market, broadband market is ADSL," observes Jay Fausch, senior director of marketing and business development at Alcatel USA. "Other forms of DSL have nice niches in various applications, but ADSL is getting a lot of attention currently. It has a fair amount of deployment activity not just in North America but around the world."

ADSL originated as a way to provide video-on-demand, he continues. "That turned out not to be the killer app that would ignite the interest of DSL," he says. "It was the explosion of the Internet, the amount of application and the content of that application that people are making available to you and me."

The mass-market appeal of ADSL will provide new revenue streams to margin-limited telcos. Because of its asymmetrical nature, ADSL is being touted as the dream service for Internet access in both residential and small business environments. Its downstream speeds can reach 7.1 Mb/s, substantially more than the upstream link of up to 640 kb/s.

ADSL was "conceived of as a way to deliver switched video service over copper because of the [large capacity] downstream and limited upstream control. It's serendipity that Internet access application lends itself to ADSL," Griffin says.

The new "splitterless" ADSL, dubbed G.Lite, was recently sanctioned by the International Telecommunication Union. This solution removes the POTS splitter from the customer premises. The splitter separates the ADSL signal from the voice signal and eliminates the need for telcos to roll a truck for each installation-clearly a big savings. G.Lite supports downstream speeds of 1.5 Mb/s and a maximum of 500 kb/s upstream. It uses less power, typically can reach longer distances than ADSL, and interoperates with ADSL. The G.Lite standard is expected to be ratified in June 1999.

Standards deviation HDSL has been around for many years, and about a year ago manufacturers convened to upgrade the standard from a single-pair to a two-pair solution. The group has since reached a consensus, and Griffin expects that the new standard-HDSL2-should be signed within six months. In the meantime, "the silicon [based on HDSL2] is being implemented by multiple players as we speak," he says.

The process was simple, Griffin says. "There was no forum. The participants were basically the people selling HDSL."

But defining an ADSL standard hasn't been as smooth. The Universal ADSL Working Group has many more participants, impeding advancement, he notes. "It's harder to get a consensus with 20 or 30 people than if you have five in a room," he says, adding that the group is "frustrated with the progress" of the ADSL standard.

Part of the confusion stems from the debate between carrierless amplitude/phase modulation (CAP) and discrete multitone (DMT). Both are line codes that enable DSL, and vendors were split on which to use. CAP was a lower-power, lower-cost solution, but DMT performed better.

The ADSL Forum, which offers its recommendations to the ITU, hosted the CAP vs. DMT Olympics in 1993. DMT won. That set the standards process in motion, says Bill Rodey, vice chairman and treasurer of the ADSL Forum, and vice president of sales and marketing at Westell Inc., speaking on behalf of the ADSL Forum.

"We are prioritizing our work on DMT. Members can propose work [on other line codes] and volunteers will look at it, but it's not a priority," Rodey says. The ADSL Forum focuses on Layers 2 and 3, so it is "line code-agnostic," he adds. The group concentrates on the framing and connection layers. In 1999, the forum will expand its scope to address the service architecture.

Marc Zionts, CEO of Westell, brushes off the CAP vs. DMT debate. Westell's approach is to focus on the application, he says. "It's about the service and provisioning as opposed to an algorithm. We have DMT and we have CAP. We started with CAP, but we'll end up with both."

The standards debate further clouds interoperability issues. Today, nearly all DSL implementations require the same vendor's equipment at the customer premises and in the CO or DLC. With the adoption of DMT, efforts are underway to improve interoperability, but that is a few years off.

The trouble with mass marketing Interoperability is only one hurdle DSL must overcome to become ubiquitous. Another is that the carrier business model-and the technology itself-seem to be squelching swift adoption of DSL, even as customers seek high-speed access.

"The demand for DSL is driven by the demand for higher-speed services at lower costs," says Paradyne's Weiner. "[But] it's a challenge for the mass market. [Telecom companies] work better when everything is vanilla."

Weiner compiled a series of issues that are hindering mass deployment of DSL. Topping the list is the need for splitterless operation. To have mass-market appeal, the DSL modem has to work like a dial-up modem that can be plugged in without a technician.

The Bell companies project that 200,000 lines will be installed this year, Weiner says. "That's great business-more than 1000 lines per day. But if you have to roll a truck to every home, and each [service call] takes one to three hours, how many trucks to do have to have in your fleet to install 1000 lines per day?" he asks.

The solution is splitterless DSL, Weiner says. Most splitterless discussions involve in-line filters that customers could install directly on the phone. Paradyne's implementation is the Multiple Virtual Line modem.

Beth Gage, senior broadband consultant with TeleChoice, agrees. "Splitterless [DSL] should be a show-stopper," she says. It will make DSL more cost-effective and push the technology forward.

G.Lite has its own detractions, however. One is that the most recent G.Lite proposal is based on Category 3 premises wiring, but home wiring varies. In addition, in-line filters can't be installed on more than three phones per home, and quality varies depending on the type of phone.

Another challenge Weiner notes is the trade-off between loop reach and speed. The farther the distance, the slower the DSL service. ADSL doesn't work past 18,000 feet, and customers at the outer edge of that distance will get limited speed. Service providers have devised tricks to extend the reach, but in some cases, DSL simply isn't an option.

To accommodate customers all over the service region, Covad Communications offers multiple versions of DSL. Lou Pelosi, director of marketing at Covad, claims his company's approach reaches most customers (see sidebar on page 32).

Speed is a lesser issue, says Claudia Bacco, senior DSL analyst for TeleChoice. Although 7 Mb/s ADSL service is possible, it's not necessary today.

When Bell Atlantic announced it was offering a 7.1 Mb/s ADSL service, some analysts said that was overkill. "What are you going to do with it?" Bacco asks. "What applications on the Internet support 7 Mb/s? People need to understand that you can buy 7 Mb/s DSL, but if the access server on the Internet supports 128 kb/s, that's as fast as you are going to go. Most people would be hard-pressed to find anything in that bandwidth." In situations in which the carrier controls the backbone from end to end, such as remote access to a LAN, you might be able to use the full 7 Mb/s, Bacco says.

Other deployment hurdles exist, according to Wiener. He cites spectral compatibility (which can affect loop reach and speed), the ability to test and qualify lines, the increasing use of DLCs (which require line cards to terminate DSL before backhauling data), power requirements and port density, among other concerns.

TeleChoice's Gage claims that many of Weiner's points are minor issues today, and carriers and vendors are addressing these for future product incarnations.

Westell's Zionts agrees. Every technology has its limitations, he says. "You can serve 90% to 95% of the market with DSL-based services. That's nothing to abandon ship for," he says.

Competition: cable and ISDN While the DSL issues are under discussion, competitive technologies such as cable modems are gaining ground. In fact, cable companies are forcing telcos to respond, Zionts says. By not offering high-speed access, telecom companies are losing customers to cable companies.

"If phone companies had the choice to roll out DSL, they wouldn't do it," Zionts says. "For cable companies, data is the entry point. They don't have a choice anymore. Now it's a matter of the phone companies doing it."

Cable modems are well-suited for residential use-for a single user connecting to a single endpoint. Detractors note that cable modem communications aren't secure. A neighbor could hack into the network and capture your bank account information if you were banking on-line, for example. And because the line is shared, the available bandwidth depends on how many of your neighbors are on-line.

DSL offers a dedicated connection, increasing security and guaranteeing bandwidth. Even if your entire neighborhood signs up for DSL, your service is unaffected. DSL also simultaneously supports multiple sessions, says Zionts. "With DSL, you can connect multiple PCs at the same time but to different destinations."

ISDN, the telecom industry's first answer to high-speed access, was largely ineffective for mass-market use. When ISDN hit the market, it wallowed as telcos tried to figure out what to do with it. Telephone companies failed to establish marketing programs, and ISDN inquiries often fell through the cracks. It was expensive and inefficient, too.

"ISDN exacerbated congestion on the network because ISDN would nail up the connection to the switch," says Zionts. "Everything is different now. ISDN was searching for an application. Today we have applications searching for technology. [With] electronic commerce, Internet growth, telecommuting, voice over IP and video over IP, the need for speed is there."

Competition isn't limited to technology, however. With deregulation, service providers are constantly seeking new services to differentiate their offerings. The risk-taking competitive local exchange carriers are often the first to make DSL available in a region.

"CLECs spur people to act," Zionts says. Although CLECs are quick to define and implement new technologies, incumbents will fight for their share, even if they enter the market late.

Competition for access will stimulate innovation, Zionts says.

"Bandwidth is like water: It will find its path," he says. "It's not a matter of if; it's a matter of who gets what share. It's a damn big market, and there is room for [many technologies], but I strongly believe DSL will get more than its fair share."

Now DSL is at a crossroads. From the carrier standpoint, DSL is the best solution until fiber to the home becomes a reality. But it has stiff competition. Service providers are competing with each other and with other technologies. DSL has to catch on, or carriers will miss the boat.

Covad Communications has been in the DSL arena since December 1997. The Santa Clara, Calif.-based competitive local exchange carrier saw the DSL niche and went for broke. About 10 months after opening its doors for business, Covad had deployed service in four metro markets, and the "packet CLEC" has aggressive plans for the future.

Why DSL? It's fast, it's here and it's ready for deployment, says Lou Pelosi, director of marketing. "DSL services are cost-effective today as a high-speed data service. With cable modems, you have to upgrade an entire network to serve one customer," he says. DSL requires only a "rack and a cage" in each central office and "the majority of the cost is borne when you have a subscriber," he adds.

Part of Covad's plan is to go after specific market segments. The service provider targets corporations and teleworkers, whose bills are paid by the company. The situation is ideal for telecommuters. "They get fast, always-on access to the corporate network," Pelosi points out. "They don't have to dial up and download [files]. A big attachment just zips down [the wire]."

As important as the speed is DSL's always-on connection. "That's very useful for e-mail, advanced Web applications and access to the corporate database," he says.

Although asymmetrical DSL supports voice and data on the same line, Covad runs only data over the line. Pelosi says. Two factors led Covad to that determination. First, most corporations will pay for their home workers to have two lines: one for voice, one for data. Second, Covad doesn't have to install a splitter at the home, so they call roll out service more quickly and save expensive tech time.

Internet service providers are another target market. ISPs can bundle Internet services with the DSL service and sell the packaged solution to small businesses.

Because ADSL only works for about 70% of Covad's service area, the CLEC also supports IDSN DSL and symmetrical DSL. "ISPs don't want to turn away three out of every 10 calls,"Pelosi says.

With the combination of services, "we have blanket coverage," he adds. "Nearly 100% [of all potential customers] are served out of the COs we're in."

Today Covad offers service in the San Francisco Bay area and in the greater Los Angeles area, as well as Boston and New York. The CLEC plans to deploy service in 18 additional markets next year. The company hopes to bring service to 5 million homes and businesses this year and 20 million next year.

Gulftel Internet, a division of Gulftel Communications Inc., has identified symmetrical digital subscriber line technology as a cost-effective way to take advantage of its embedded copper infrastructure to provide customers with high-speed local loop access.

Founded in 1995, Gulftel Internet serves a winter resort area along the Gulf coast of southern Alabama and has 7000 dial-up subscribers and about 250 commercial business subscribers. Deploying TUT Systems' Expresso DSL access multiplexers (DSLAMs) within a five mile radius of its central office, the company was able to use existing copper to provision commercial subscribers with up to 768 kb/s Internet access.

There are three remote Internet service provider offices where the TUT System Expresso DSLAM terminals are deployed to hub SDSL local users. SDSL technology has also allowed Gulftel to provide dedicated high-speed connectivity for an Internet-based hotel and resort reservation service.

Another one of Gulftel's remote offices provides 768 kb/s local loop connectivity for a golf reservation system linking eight golf courses via SDSL and frame relay. During the winter season, club members at any one of the golf courses can reserve a tee time at any one of the courses.

The third remote office is used to support Gulftel's remaining commercial and dial-up subscribers.

The Expresso GS DSLAM terminals are all linked back to the CO over dedicated T-1 links, which can be upgraded to T-3 connections.

Service providers that deploy SDSL services using "dry" copper should be aware that in some serving areas they should establish a test circuit before extending service to the end user. This precaution will allow the ISP to determine if the local serving office has placed filters on the lines. Filtering may protect adjacent pairs in a service bundle, which can affect service speed. Some carriers may deploy line filters to discourage the lease and release of dry circuits for high-speed access. ISPs in various parts of the country have reported some problems with their local utilities and filtered lines. In some cases, the utility claimed that its service was performing to published specification.

In January U S West moved into the digital subscriber line game by announcing an aggressive deployment plan, and the Bell company has been working on that plan for the better part of the year.

In the second quarter of next year, U S West will phase DSL service into more than 220 wire centers across 40 cities, says Mike Rouleau, vice president of marketing at U S West !nterprise Networking. "Our deployment is aggressive in rural and metro areas, and we will continue to expand that going forward," he says. "We see opportunities for more cities, and there is great opportunity to build out the networks where we are already."

U S West's switch-based network is designed to support the boom in data services. "The market opportunity for Internet access demands a switched network infrastructure as opposed to a routed one, and a switched network moves packets faster and more reliably," Rouleau says.

The carrier focuses on providing fast Internet access and LAN-like capabilities to the home, he says. "Today we offer full-rate DSL technology. We deliver [symmetrical] service at 256, 512 and 768 kb/s and 1 Mb/s, and we offer ADSL at 4 Mb/s downstream and 1 Mb/s upstream or 7 Mb/s downstream and 1 Mb/s upstream."

The speed flexibility is geared toward various teleworker applications, but U S West is focusing on the application as well asthe technology, Rouleau points out. "It doesn't mean anything to a customer to say, 'I can sell you RADSL. [They want to hear] 'I can offer you Internet access and LAN-like service,'" he says. "We believe DSL is the platform for Web tone. It will enable customers to take advantage of Web-based applications and services, and it's scalable."