can ip answer at&t's call?
May 26, 1998
InfoWorld via NewsEdge Corporation :
Faced with
threats to its markets from all sides -- including the
new breed of carriers such as WorldCom, the
emergence of Internet telephony, and a shake-up of
the regulatory landscape following the 1996
Telecommunications Act -- granddaddy of the
telecommunications industry AT &T is throwing all of
its weight behind IP technology. But the company's
long-term plan to migrate its entire network
infrastructure to IP is winning mixed reviews, and
many see it as risky.
AT&T divulged its IP infrastructure plans earlier this
month at the opening of its Silicon Valley
headquarters for AT&T Labs, in Menlo Park, Calif.
(See "AT&T plans migration to IP," May 18, page
55.) The company intends, over the long-term, to
move its entire network infrastructure -- including its
long-distance voice network -- to an IP platform
called the Advanced Network Services Platform that
will serve as the simplified basis for the creation of
future services for AT&T.
Generally, analysts had two concerns about the
plan: that the quality of voice calls over IP is not high
enough to satisfy users' expectations, and that IP is
largely untested in such an environment.
"Putting all on a single infrastructure -- Internet,
virtual private networks, voice needs -- that's
potential for a major hit," said Tom Jenkins, an
analyst at TeleChoice, in Verona, N.J. "The [voice]
quality issue will go away within the next 12
months, but the question that will remain is
reliability."
The reliability issue is seen as particularly
significant in the light of the AT&T frame-relay
network outage in April. (See "AT&T disconnects, "
April 20, page 1.) But analysts differed on whether
the risks would have been increased or decreased, if
it had been an IP network that was involved.
"Because IP travels with self-healing routing, they
won't be vulnerable to cut off; if packets get blocked
on the way, they'll find another way to the
destination," said John Nitzke, an analyst at
Forrester Research, in Cambridge, Mass.
However, others were more cautious.
Brett Azuma, an analyst at Dataquest, in San Jose,
Calif., said a broadcast network storm, which
caused the frame-relay outage, is also possible with
an IP network. And IP networks can have more
points of failure, he added, because IP networks
tend to try to squeeze more capacity into smaller
boxes.
Overall, analysts said much will depend on AT&T's
implementation.
"Network design has a lot to do with how
failure-proof it is," said Hilary Mine, an analyst at
Probe Research, in Cedar Knolls, N.J. "It has to do
with how you build redundancy, if there are
[Synchronous Optical Network] rings, if there are
dual processors, how you route traffic."
AT&T admits that there is a long way to go before
implementation.
"This will take a massive amount of work," said
Audrey Curtis, development vice president of AT&T
Labs. "We need to ensure that the underlying
backbone has the capacity, the router infrastructure
is robust, and that the gateways allow a multitude of
services to flow into the backbone, and that they're
also scalable and robust." [Nice opportunity here]
"A lot of boxes need to be invented," agreed Frank
Dzubeck, president of Communications Network
Architect, in Washington. "And [AT&T] just turned
up its first voice gateway in trials, and it takes 28
numbers to dial. "
Implementation aside, some analysts suggested
that AT&T has little choice but to embrace IP.
"[AT&T] understands the threat of IP networks, and
so would rather ride the wave than get crushed by
it," Azuma said.
Jeff Pulver, president of pulver.com, an Internet
telephony research company, added that AT&T is
intelligently betting on the belief in the industry that
public networks will be IP-based.
"[AT&T] has identified that IP telephony is definitely
the third wave of communication behind the
telephone and cellular," Pulver said. " They're going
with technology that will be there in the future -- the
biggest revenue opportunity for the 21st century."
Analysts also pointed out the network efficiencies
that AT&T will receive by consolidating to one
network: Not only will there be a single network to
manage, but there will be a better use of bandwidth,
using voice compression. By putting voice into
packets, Nitzke said, the carrier will receive a 50
percent savings in bandwidth, because of more
efficient bandwidth utilization.
But despite AT&T's good intentions, analysts
question whether it's just that - - a good intention.
"The question is, is the timing right?" said
TeleChoice's Jenkins. "Gateway devices and servers
are new, unproved, and so for AT&T this move may
mean potential high rewards, but also high risk."
AT&T's move to IP is the culmination of various
activities by the company in this space. Prior to the
opening of AT&T Labs, the carrier made headway in
the voice-over-IP arena with its WorldNet Voice --
now called AT&T Connect 'N Save Service -- IP
telephony offering (see "AT&T's next voice," Feb. 9,
page 1), and its Global Clearinghouse (see "AT&T
service acts as broker," April 13, page 10).
Meanwhile, MCI and Sprint are not on the same
track as AT&T. An MCI representative noted that
although IP is the wave of the future, the company
has no great urgency to turn its network to IP.
However, she added that MCI's Vault Initiative, which
bridges the Internet and the switched network, is the
direction the company is taking with IP. (See "MCI
lets Net users click for customer service," Feb. 2,
page 42.)
Sprint, on the other hand, is committed to ATM,
according to a representative.
Ultimately, whoever comes up with the most winning
strategy stands to dominate wide-area networking
well into the next century.
AT&T Corp., in Basking Ridge, N.J., is at
att.com. Sprint Corp., in Kansas City,
Mo., is at sprint.com. MCI
Communications Corp., in Washington, is at
mci.com.
[Copyright 1998, InfoWorld]
Copyright c 1998, NewsEdge Corporation No redistribution allowed. |
