SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Ascend Communications (ASND) -- Ignore unavailable to you. Want to Upgrade?


To: djane who wrote (52225)8/18/1998 2:22:00 AM
From: djane  Respond to of 61433
 
Packet technologies force telcos and vendors to re-evaluate the circuit switch [Good ASND and LU references at the bottom]

internettelephony.com

Cover Story

August 17, 1998

FROM THE GROUND UP


VINCE VITTORE and SANDRA GUY

When Sprint announced in June that it would build a new network based on packet
and cell technologies, boosters used words such as "bombshell," "revolutionary" and
"a watershed event." In many ways, the adjectives were correct. Never before had
a carrier come forward with a plan that would propose to change the face of the
network as greatly as the Integrated On-Demand Network.

Analysts, however, viewed the plan as nothing new--simply the culmination of
high-speed access events that have been building for five, even 10 years. Yet
Sprint's proposal may provide a glimmer of the inner workings of the unfolding
central office (Figures 1 and 2).

Unlike today's CO, in which telcos must tweak voice switches to carry data traffic,
the CO of the future likely will look much more like a collection of large data
switching centers.

Other options are emerging, as well. Level 3, a new Omaha-based carrier that
expects to turn up Internet protocol (IP)-based service in the third quarter, uses no
circuit switches. When it deals with incumbent LECs' COs, the carrier hopes to
leverage unbundled elements, says Jack Waters, vice president of network
engineering at Level 3. And in metropolitan areas, Level 3 will build large gateway
nodes and local networks that touch end user buildings and Internet interexchange
points.

"Our local ring, instead of touching COs, touches different kinds of facilities,"
Waters says. Level 3 describes itself as a communications and information services
company that provides a full range of services, including local, long-distance,
Internet and other enhanced services over IP.

Yet the process of using unbundled network elements remains uncertain because
the issue is tied up in court. That's where the regulatory stumbling blocks come in.
Transport-related data services such as ATM and frame relay are regulated, but IP
services are not.

John Charters, vice president of Internet service for U S West, wonders whether
regional Bell operating companies' Internet service providers will be motivated to
push convergence in the CO. In this regulatory climate, RBOCs are motivated to
push convergence into the unregulated IP point of presence (POP) created to
deliver IP services.

U S West has petitioned the FCC to deregulate all data services. However, the
FCC recently recommended that incumbent carriers be exempted from unbundling
or reselling elements of their data networks to competitors.

At the end of the network, things also will look different. Instead of end points that
do nothing more than pass on information, the network of the future will collect
information at the premises, process it at a service node and add intelligence.
Additionally, network intelligence, which has traditionally resided deep in the
network where telcos could easily maintain and upgrade it, will be pushed out in the
form of applications, which could be stored locally, domestically or internationally,
says Marty Kaplan, senior vice president and chief technology officer for Sprint.

Indeed, the entire idea of having a CO could undergo a radical change, says
Warner Andrews, strategic product line manager for ATM at Rockwell
Semiconductor Systems. "The concept of a CO becomes a collection of control
points for cell traffic, as opposed to a big box [or five-story building]."

Andrews calls Sprint's ION announcement "a harbinger of good things to come for
people who supply ATM gear." He also perceives ION as an example of "using an
ATM protocol to obliterate the concept of the CO."


Others call the future CO an edge vehicle that will serve a public network that will
look much like today's enterprise network--multivendor, multiprotocol and
multinetwork management, says Hilary Mine, senior vice president at Probe
Research. "The good news is that, for the first time, service providers will call the
shots. The bad news is that the network will be messy and difficult to monitor,
maintain, operate and provision," she says.

OSS effects

The idea of using IP to push voice through is even having its effect deep within the
traditional CO, where operations support systems (OSSs) have hummed along for
years in the circuit-switched world. For carriers such as Sprint and Bell Atlantic,
which are both experimenting with IP voice services, the transition means
implementing new software and systems. Just as important is finding the manpower
to run them.

"We believe in six key business processes--billing, customer care, provisioning,
trouble administration, network engineering and network management," says Jason
Donahue, vice president of marketing for Beechwood, a Clark, N.J.-based
consulting group specializing in OSS integration. "What IP telephony is doing from
an OSS perspective is changing the way the carrier manages those processes. What
that means for telco business processes is the systems themselves are going to
become more complex as the network elements mature."

Donahue uses billing as a prime example of the change IP services will have on
traditional OSSs. Because carriers will be able to provide multiple services on one
platform, customers will begin demanding integrated bills. "Customers want a single
bill, and they want a discount. To achieve that you don't need a single rating engine,
but at the very least you have to have integration between different billing systems,"
he says.

Incumbent carriers will need to install new support systems to handle IP-based
services. The trick over time, he adds, is to create a common set of plans, and that
may not happen for four or five years.

Economics

Regardless of where carriers are in their IP plans, one simple truth remains: The
potential profits from IP are too compelling to ignore.

The fact that data is catching and sometimes surpassing voice traffic on the public
switched network makes an economic case for switching to IP networks, says Ray
Keneipp, principal analyst for carrier infrastructure at Current Analysis.

The addition of intelligence on the end of the network strengthens that plan, says
Mine. In the end, she sees a mix of models not unlike the enterprise world today. In
most areas, carriers will have a combination of circuit switches and IP data
switches.

However, not everyone is convinced. Fred Seigneur, chief technology officer of
Convergent Communications and president of Sonetech, a think tank/systems
integration company, says the cost of building IP networks is higher than many think
because voice compression over the Internet requires an expensive digital signal
processor on each end of the call. Convergent, which bills itself as an enterprise
carrier, is using a client/server architecture called ROADS, which stands for robust
open architecture distributed switching, to provide customers with an integrated
voice/data service. That won't necessarily mean IP-based voice, but it won't
preclude it.

But Level 3's Waters says that the carrier uses no voice compression. "We're in the
middle of a development plan to build a substantial fiber optic channel.
Compressing 56K voice channels is not at all the end game we're shooting for," he
says.

Ironically, the economics driving carriers in the IP world appear to work in reverse
for the OSS portion of the network, he says. As carriers gain the operating
efficiencies from IP, OSS becomes more important as a differentiation and more
expensive in terms of percentage of capital budgets.

Although some OSS vendors are trying to develop standardized products for the IP
environment, Donahue brushes off the idea as unfeasible, particularly for new
entrants such as Qwest and Level 3. "In a sense, the new carrier is more complex
because it has to learn the business processes first, and there is no telco in a box,"
he says. "It's virtually impossible to create off-the-shelf OSSs given the complexity
of the telco environment. What the bigger carriers learn as they roll out services is
that their business is different from any other carrier's."

The vendor angle

While network operators continue to push for singular platforms, the shift to an
IP-based environment is causing cataclysmic changes in the traditional vendor
world. Big switch manufacturers such as Lucent Technologies, Northern Telecom,
Ericsson and Siemens Telecom Networks can no longer stake their futures on a
steady flow of multiyear contracts from telcos. The rapid ascension of IP's
importance in the CO is forcing them to reposition themselves, either through
internally developed IP programs or acquisition. [Hmmm....]

More important, the CO market is opening more to companies such as Cisco Systems, 3Com and Ascend Communications that have years of IP experience.


"The voice guys say, 'We can buy our way into this,'" says U S West's Charters.
"And Cisco people say, 'We can engineer our way into this.'"

The question is not so much what happens to the CO, he adds, but whose
equipment goes in it. Regardless of the vendor name on the CO boxes, he doesn't
think anyone has yet presented a unified voice/data POP. "The interaction between
my data tone and my dial tone is occurring. The jury is still out, in my mind, as to
what needs to occur," he says.

Vendors, however, don't have the luxury of waiting and must place their bets now.
Among the stream of announcements earlier this year was one from Nortel that
perhaps sums up the position big vendors are in better than any other. In June, the
company agreed to acquire Bay Networks, giving it a strong player in the IP market
and what some thought would be an entry into the voice-over-IP world. After all,
Bay already had pushed voice over its network equipment and through an earlier
acquisition of LAN City was among the forerunners in the cable modem market.
Still, less than two weeks after the acquisition, Nortel unveiled its own IP strategy
called IPConnect (Figure 3).

Geared for all levels of carriers, the IPConnect line runs from a low-end 24-port
product that could be integrated with PBXs up to Nortel's plans to integrate IP
ports on its DMS switches. Throughout the whole line, the company is looking at IP
voice as a business tool, which requires some different service assumptions from
those that most gateway vendors are under.

"What we're saying is if you're interested in toll arbitrage or getting out of access
charges, don't do IP telephony," says Al Bender, vice president and general
manager of carrier voice-over-IP solutions for Nortel. "IP is the holy grail that
industry has been searching for, which is the ability to integrate voice and data over
the same infrastructure. But before IP telephony can be a serious business tool, we
believe there can be no change in behavior."

And although that will initially mean providing single-stage dialing (as opposed to the
two-stage calling card method most operators use today), it also means providing
the type of applications users now take for granted, including network-based voice
mail, caller ID, three-way calling and call waiting.

To do that, vendors also are being forced to do something they've never had to
do--open their switches to third-party vendors. In the circuit-switched world, any
new application development generally is handled by the vendors themselves or a
small circle of software companies. The advent of IP not only is forcing vendors to
develop open interfaces to their switches, it's creating some unique opportunities
and services.

Lucent, which plans to integrate some of the capabilities of its PacketStar IP switch
into its 5ESS, will use a combination of internal development and third-party
vendors, says Dianne Herr, vice president of marketing for the 5ESS group. And
because applications likely will sit out further in the network, the company is looking
to work closely with edge device vendors.

"I think it's critical that we have that gateway function in the 5E and as a separate
product," says Herr. "People can develop these capabilities and utilize them with
each other."


"In the Internet space, everyone has learned that they can't do everything alone,"
says Graham Howard, manager of advanced applications for Siemens. "We want to
build an enduring architecture that's not just tied to technology today."

Any Comments?
Send them to Karen Murphy at msblues@earthlink.net.

www.internettelephony.com
Telephony August 17
c1998 Primedia Intertec
All Rights Reserved.

| HOME | JOB ZONE | SUBSCRIPTIONS | ARCHIVES |