Chicago -- In spite of ongoing public announcements about
wide rollouts, telcos' digital-subscriber-line-based service deployments are being slowed
by confusion over pricing, by a lack of standards and by an alphabet-soup of services from
which to choose.
That was the clear message during a recent two-day seminar
here on xDSL hosted by the International Engineering Consortium, which drew about 250
The conference focused on DSL technology, which allows
network providers to transport high-speed data over existing copper telephone lines. It is
widely considered to be the technology of choice when it comes to competing with
high-speed cable modems.
But generally, telephone companies are running blind in
their attempts to roll out DSL, in part because the technology is limited by the distance
and quality of the existing copper lines. This means
that some customers -- namely those who reside closer to central offices, or those served
by high-quality circuits -- could actually take advantage of higher bit-rates than others
who aren't as conveniently located.
Consequently, telcos are unsure about which data rate suits
them best. Because of this disparity, they have developed a wide variety of bit-rates.
For example, the best-known service, asymmetrical DSL, can
transport data downstream at 6 megabits per second, but it offers just 640 kilobits per
second in the upstream direction.
Other flavors of DSL exist, however. Symmetric DSL operates
at 1 mbps in both directions, which makes it ideal for videoconferencing and other
applications that eat up bandwidth symmetrically; HDSL (high-bit-rate DSL) works at T-1
rates for ever-faster communications, but it requires two copper pairs; and VDSL (very
high-bit-rate DSL) pumps data downstream at 51 mbps and upstream at 3 mbps.
Compounding the confusion over which version to offer is
the lack of completed standards that would allow hardware from different vendors to
This gap puts the DSL marketplace behind the cable
industry's DOCSIS (Data Over Cable Service/Interoperability Specification) effort,
according to Jeffrey Goldthorp, general manager and senior director of broadband-access
design and engineering at Bellcore, the telephone industry's research-and-development
"The DOCSIS specification is evidence that there's
common ground and cohesion in the cable industry," Goldthorp said. "The ADSL
movement is more fractionalized."
He added that an interoperability test bed, like the one
that Cable Television Laboratories Inc. has set up, would be "an important step"
toward making DSL equipment that plugs and plays over any telecommunications network.
Despite that hurdle, the telcos could quickly catch up to
the cable industry, in part because at least 60 million copper circuits are both short
enough and of sufficient quality to handle the service, Goldthorp said.
Conversely, the cable industry is only now upgrading its
networks to support two-way signal transport, and it will be at least three years before
it's substantially finished, he said.
Other issues that threaten to hamper DSL deployments
include: disagreement over whether the "core" network should be based on ATM
(asynchronous transfer mode), IP (Internet protocol), or a combination of the two
solutions; a lack of network-management software; the relatively high cost of the modems;
and possible cannibalization of the telcos' existing high-margin T-1 business.
Technology aside, one of the biggest issues facing telcos
is pricing. Companies that are accustomed to charging several hundred dollars per month
for T-1 lines will be loathe to give up that lucrative business, although they may have to
do so in order to compete directly with cable modems.
Still, the technology works, and subscribers enjoy its
benefits, according to Charles Gowder, general manager of the Valley Telephone
Cooperative. Gowder's telephone system serves 5,600 people via 17 exchanges in rural south
Texas, and he recently connected his 16th DSL subscriber.
Valley chose to deploy the service in these less-than-ideal
circumstances because it wanted to be perceived as a leader, Gowder told conference
attendees. The past year has been consumed evaluating hardware, performing a cost study
and testing the service with three customers and one school, he added.
Gowder's company charges $36 to condition the line for the
service; $30 for basic Internet access; and $70 more for basic installation ($225 for full
turnkey installation, including an Ethernet line card), and it is currently losing about
$4,000 per month, he reported.
Although it was far from passing the market test of
viability, Gowder was upbeat about the technology.
"I just wanted you to know how this operates in a
worst-case scenario," Gowder said jokingly to the primarily telco-based audience.
Along the way, Valley learned several painful lessons,
including how to design the network for optimum performance and how to deal with customers
who could not get service because they lacked the necessary Ethernet cards.
But perhaps the biggest hurdle to deploying services to the
mass market is a lack of network-management hardware and software to serve the large
carriers, said Shawn Shokoohi, senior member of the technical staff at BellSouth Science
Several computer companies touted their management systems,
"but they don't know telecom," Shokoohi said. "We don't know who to go
to," he confessed, noting that although Lucent Technologies and Bellcore offer such
software, it's still evolving, and it doesn't offer a consistent set of features.