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: Greg Jung who wrote (46861)5/14/1998 2:40:00 AM
From: djane  Respond to of 61433
 
MY Times. Bay Vaults on a Takeover Story [Info on NT share dilution]

BY SETH SCHIESEL, May 14, 1998

There are not many rock-solid articles of faith about the future of
the technology business, but if any belief comes close, it is that
companies that make telecommunications equipment and companies
that make data-networking equipment are destined to merge.


That is why a news report that Bay Networks Inc. (a data-networking
company) had rejected a takeover offer from Northern Telecom of
Canada (a telecommunications equipment company) touched such a
nerve on Wall Street Wednesday, sending Bay's shares up $3.9375, or
16 percent, to $27.9375 on the New York Stock Exchange.

After all, the thinking went, if Bay had spurned Northern Telecom, one
of the world's largest makers of the big computer boxes used in
telephone systems, a better deal must surely be in the works.

A person close to Northern Telecom said Wednesday that the two
companies had discussed a merger, though the person would not
confirm that a specific offer had been made.


Wednesday's report from Bloomberg News had Bay, and perhaps
even the company's chief executive, David House, telling Wall Street
analysts at a meeting last week that Bay had rejected an offer from
Northern Telecom as paltry.

"That's not accurate," said Mike Deshaies, a Bay spokesman. "What
they're saying about what we supposedly said is wrong."

Northern Telecom refused to comment Wednesday.

Tim Luke, a research analyst for Lehman Brothers who follows Bay,
said he did not receive a message about a Northern Telecom bid from
House. "I sat through the analyst meeting and he certainly didn't say
anything like that," Luke said. "What he did say is that it will be up to
shareholders to decide what the best value is."

But Luke's wait-and-see attitude was not in vogue Wednesday. Most
traders seemed to think a bit more, like Michael Cristinziano, a
high-technology analyst for Gerard Klauer Mattison. "Did I hear it from
him directly? No," he said of any comments House may have made
about an offer from Northern Telecom. "But do I believe it makes
complete sense that Nortel would want to buy a data networking
company like Bay? Yes."

Many people believe that many companies would want to buy a
networking company like Bay Networks. Lucent Technologies comes to mind.

But Lucent and Northern Telecom have accounting and ownership issues that make their pursuit of an acquisition more complicated than it might appear.

Lucent cannot take advantage of a mergers accounting practice known as a pooling of interests until this October. In a pooling, a company that
is buying another company for stock avoids having to take charges
against its future earnings to cover the difference between what it paid
and the assessed value of the target company's assets.

Lucent was split off from AT&T on Oct. 1, 1996, and accounting rules
prevent the company from pooling until two years from then.

Lucent used cash last month when it agreed to acquire Yurie Systems
for $1 billion. But when the potential target has a market value
exceeding $6 billion, as Bay does, pooling and waiting until October
may begin to appear more attractive.

One of the Machiavellian possibilities swirling around Wall Street Wednesday had Northern Telecom sniffing around Bay Networks just to drive up Bay's stock price and perhaps stampede Lucent into acquiring Bay with stock before it could use the pooling method (thus heavily diluting Lucent's future earnings).

That is a bit of a stretch, especially since Northern Telecom could have problems of its own making an acquisition with stock.

BCE Inc., a holding company that owns just over half of Northern Telecom, has said it will not allow its stake to be diluted below 50 percent. Were Northern Telecom to issue $7 billion or $8 billion in stock to make an acquisition, BCE's stake would surely fall below that level. To compensate, Northern Telecom would have to issue even more stock to BCE. That could be intolerable to other shareholders. So Northern Telecom would probably have to make an acquisition with cash.

All such hypotheticals did little to deter investors' interest in the story
Wednesday, especially since Northern Telecom did so little to resolve
questions.

But perhaps it would not have mattered if it had.

After speculation emerged Wednesday that Siemens of Germany would
offer to acquire Motorola, a Siemens spokeswoman said, "We are not
interested in acquiring Motorola."

No matter. Motorola's shares rose $3.50, to $59.375.

Home | Sections | Contents | Search | Forums | Help

Copyright 1998 The New York Times Company



To: Greg Jung who wrote (46861)5/14/1998 2:46:00 AM
From: djane  Respond to of 61433
 
thestreet.com. A Buyout Offers Bay a Way Out [ASND reference]

thestreet.com

By Kevin Petrie
Staff Reporter
5/13/98 4:42 PM ET

It's not just the arbs who hope Bay Networks
(BAY:NYSE) is for sale -- some on Wall Street say it's
Bay's best hope.

Sell-side analysts and investors are scratching their
heads about what Bay executives actually said about
takeover talks early Tuesday at a trade show in Las
Vegas. Some people who attended the meeting say Bay
executives didn't say the company had turned down an
offer from Northern Telecom (NT:NYSE). Bloomberg
reported early Wednesday that Bay told analysts just
that, but that it would consider other offers.

Bay's stock reacted to the news, soaring 3 15/16 to 27
15/16 on heavy volume by late afternoon.

Wall Street's consensus is that it selling is Bay's only
way out. Otherwise, it will remain mired in the computer
networking industry's second tier. The attraction is that
Bay offers a way into the data networking market for
corporations.


That's a far cry from the optimism that greeted CEO Dave
House, an alumnus of Intel (INTC:Nasdaq), when he took
over Bay in October 1996. House swiftly patched Bay's
open sores by stanching the brain drain, clearing R&D
clutter and getting a few strong products out the door.
Operating profits started to rebound last spring, and
investors pushed up Bay's stock 160% from late April to
mid-October.

The press gushed. A Business Week story said in
headlining a story, "Mr. House Finds His Fixer-Upper."

But sales slowed early this year, across the product line,
and price competition gouged profits. In an interview early
February, House seemed tired as he voiced caution about
the quarter's outlook. Today, the turnaround is on the
rocks, as evidenced by the past quarter's poor profits.

So as an exit strategy Bay reportedly wants to sell itself
for 32 to 35 a share, or more than three times trailing
revenue, according to one trader. That's still a healthy
premium after the stock's rise today. One trader said
enough arbs piled into the stock to keep the price aloft for
a few days.

"It's a perfect play for anyone who wants to become, overnight, a player in the incumbent internetworking space," says Greg Rossman, principal at Broadview Associates investment banking firm. Rossman adds that House must either execute several strong quarters, or
consider a partner. His firm advised Bay's recent
investment in NetSpeak (NSPK:Nasdaq) and its
acquisition of Xylogics in 1995; Rossman declined to say
whether he's advising Bay now.

How did Bay fall this far?

"I don't think it's fair to pin any blame on Dave House,"
says analyst Amar Senan at Volpe Brown Whelan,
which hasn't participated in any recent Bay underwriting
projects. Senan says Bay is locked in the second tier along with 3Com (COMS:Nasdaq) and Cabletron
(CS:NYSE).


"These are companies that have succeeded in the past
because of some new hot box," Senan says. Now, "I don't
think hot boxes alone will do it." Now, Bay is unveiling
new Accelar switches for corporations, along with
products for so-called "virtual private networks" from a
recent acquisition. But what drags the company is old
baggage -- namely, so-called hubs that promise little
growth but still form a sizeable chunk of revenue. 3Com
and Cabletron, Senan says, have a similar quandary.

Now Bay hopes for an angel -- most likely a big supplier of phone gear that will acquire it to extend into data
networking for corporations.

"Bay, as a matter of elimination, is a most attractive
candidate," says a sell-side analyst who asked not to be
named. 3Com is still a little pricey, and Cabletron has lost
too much market share, he says. The analyst maintains
that Bay has gained market share, despite the recent
damage of a product transition and price competition.
(Ascend (ASND:Nasdaq) also is an attractive acquisition
candidate, though for a bidder seeking to bolster its
presence in the data market for phone carriers.)

The rumored acquirers are Lucent (LU:NYSE), Northern
Telecom, Ericsson (ERICY:Nasdaq) and Alcatel
(ALA:NYSE).

Northern Telecom, Ericsson and Alcatel each wants to beat giant Lucent. And there may be a window of opportunity for some of these companies: Lucent can't acquire a company using a pooling-of-interests transaction, the most desired method, until October, two years after it was spun off from AT&T (T:NYSE).

Officials at Lucent and Ericsson declined to comment. The others could not be reached for comment.

Lucent, which has resold Bay products in recent months, seems to some an unlikely acquirer.

"The relationship between Lucent and Bay, while still friendly, has cooled since they signed the deal," says the sell-side analyst. The companies were wary of sharing competitive information, which stifled the alliance. His firm hasn't performed underwriting for Bay.


Rossman at Broadview says Nortel boasts a strong
distribution channel for selling phone systems to
corporations; he finds Alcatel less likely because it focuses on wide-area networks rather than the small corporate systems on which Bay focuses.

c 1998 TheStreet.com, All Rights Reserved.



To: Greg Jung who wrote (46861)5/14/1998 2:51:00 AM
From: djane  Respond to of 61433
 
5/98 tele.com on intelligent vs. dumb networks

teledotcom.com
No Wrong Answers--Yet

By Dawn Bushaus, Internet Editor

It's easy to get network planners to offer their visions of the
next public network. The tough part comes when you try to
reconcile those visions to create a clear view of the public
network of tomorrow. Here are some of the key philosophical
questions that service providers and their suppliers are now
grappling with, along with some not-always-compatible
answers.

Should the network be intelligent, or should it just deliver fast,
dumb pipes?

* "The middle of the network should be as stone-cold stupid
as possible." --David Isenberg, consultant (Isen.com,
Westfield, N.J.) and former AT&T scientist

* "We're not looking to build a big, dumb network. We want to build one that's big and smart." --Ron Vidal, senior vice president of new ventures, Level 3 Communications Inc. (Omaha, Neb.)

* "The intelligence will be distributed." --C. Holland Taylor, CEO, USA Global Link Inc. (Fairfield, Iowa)

* "Ten to 20 years from now, there may be enough intelligence in the end points, but that's not the case today." --Mark
Wilson, director of strategic planning, Ericsson Inc. (Research
Triangle Park, N.C.)

* "I like the idea of a stupid network rather than an intelligent network." --Junichi Kishigami, vice president of technology and general manager at NTT America Inc. (New York)

Is IP switching or ATM better as a core technology?

* "We are working to build a fast, scalable, fail-safe ATM core." --Mike Grubbs, director of enterprise services, Sprint Corp.

* "IP goes everywhere. ATM doesn't." --Fred Baker,
chairman of the Internet Engineering Task Force and fellow at Cisco Systems Inc. (San Jose, Calif.)

* "IP on Sonet [Synchronous Optical Network] just isn't quite there yet." --Ron Vidal, Level 3

* "ATM gives me the tool to allocate bandwidth. I can't
segregate traffic today with IP." --Andrew Schmidt, product manager, Ameritech Corp.

* "We're agnostic. We'll be working with both technologies over the next couple years." --Jack Walters, vice president of network systems engineering, MCI Communications Corp.

Will carriers continue to accelerate the pace to deploy
packet-based architectures if IP telephony is regulated?

* "If there is regulation to slow IP telephony's momentum, the monopoly carriers will have a safe harbor. They will slow down their efforts." --Tom Evslin, chairman and CEO, ITXC Corp. (North Brunswick, N.J.)

* "We've been working toward a cell and packet
infrastructure that will support all real-time applications for the past five years. Voice over IP isn't really driving this." --Mike Grubbs, Sprint

All use of this service is subject to the Terms and Conditions of Use.
All Rights Reserved.

Copyright c 1998 tele.com, The McGraw-Hill
Companies, Inc.
All Rights Reserved.
Website designed by COMPUGRAPHIA
Home page designed by Dennis Ahlgrim.
Last Modified: 5-May-98




To: Greg Jung who wrote (46861)5/14/1998 2:56:00 AM
From: djane  Read Replies (3) | Respond to of 61433
 
5/98 tele.com. Cable Tinkers With IP Voice

Circuit-switched telephony was a no-go for
cable providers. Will VoIP finally turn the
trick?

teledotcom.com

By Carl Weinschenk
Carl Weinschenk is executive technology editor at
tele.com. His Internet address is
cweinsch@teledotcom.com.

Cable operators in the United States have long been teased by
an elusive pot of voice services gold at the end of the rainbow.

Providers were smart enough not to kid themselves about the
real value of that pot, and public posturings aside, their
involvement in circuit-switched telephony never gained all that
much momentum.

Lots of "couldn'ts" put telephony out of cable providers' reach.
Operators couldn't deal with the requirement that the public
switched network deliver lifeline services, like maintaining
power during electricity blackouts. They couldn't do the billing.
The biggest "couldn't"? They couldn't expect to capture too
much market share from incumbents that had a superior
reputation for reliability. So while U.S. cable operators made
some notable headway in the telephony business in their
operations in the United Kingdom, telephony was pretty much
a nonstarter in the States, except for some bypass services
aimed at commercial users.

IP telephony is changing all that, at least for now. Cable providers see voice over IP (VoIP) service as a way to break into the local telecom market without shouldering the heavy regulatory responsibilities and technical complexity of the public circuit-switched network, although the regulation of
VoIP could take some or even all of those advantages away.

But just as with cable's previous initiatives into the voice
business, things aren't as clear-cut as they seem. Despite the
interest of cable operators such as Cox Communications Inc.
(Atlanta) and Comcast Corp. (Philadelphia)--which expects to
run field trials early next year--no one's offering commercial
service yet. "There are a lot more questions than answers,"
says Mario Vecchi, chief technical officer for Road Runner,
the high-speed modem service of Time Warner Cable
(Stamford, Conn.). Among those questions:

* What exactly will cable VoIP service look like? Will it
simply be the voice component of a multimedia package? Will
it be robust enough to become a primary-line service, where
most people feel the real money is, or will cable providers be
content offering a cheap alternative for second- and third-line
service (see "Second Nature")?

* Will cable operators band together to offer end-to-end
VoIP long-distance networks, or will they take their
established track and develop their services independently?

* Is it necessary to focus on one or two VoIP disciplines to
make the business plan work across the industry?

* Will any regulation of VoIP ultimately squelch all these
efforts?

This is the season to ask those questions. "There has been
tremendous interest in this particular solution since last fall,"
says Jane Zeletes, vice president of marketing for Hybrid
Networks Inc. (Cupertino, Calif.), a cable and wireless
modem vendor. "We're aware of many medium to large
operators that are very interested in VoIP, domestically and
internationally." Zeletes says that Hybrid is talking to about 30
cable companies in the United States about either
circuit-switched or VoIP telephony.

TWO-WAY STREETS

As part of the general upgrade move to hybrid fiber-coaxial
infrastructure, the cable industry set baseline standards for
bidirectional high-speed data. Those standards are embodied
in Multimedia Cable Network System (MCNS), developed
by industry consortium Cable Television Laboratories Inc.
(CableLabs, Louisville, Colo.). The next piece of the
technology and deployment puzzle for VoIP and other
advanced services is to establish standards beyond MCNS
that will allow the lower latency and guaranteed levels of
service necessary for telephony. This effort, called
PacketCable, will pave the way for telephony and advanced
multimedia applications.

Voice will be an ancillary part of these services, as with audio
feeds for the videoconference services that cable providers
hope to deliver. Cable operators and the VoIP equipment
vendors that hope to do business with them will use the basic
and extended standards set during the previous two stages.
The industry is racing to set standards both for standalone
VoIP and as the audio element of the advanced packages.
"We set a spec in record time," says Steven Craddock, vice
president of new media development for Comcast and a key
player in the PacketCable efforts. Working products will be at
the Western Cable Television Show in Anaheim, Calif., in
December, and field trials will start early next year, Craddock
says.

The key question is how far the industry will go to position
VoIP as a primary-line service. This is far more complex,
demanding--and lucrative. "When you look at the initial plan,
you have to italicize the word 'initial,' " says Michael Harris,
president of Kinetic Strategies Inc. (Phoenix), which publishes
a cable industry newsletter. "It's not fully fleshed out what the
capability of the plant is."

The technical issues to the creation of a robust platform
capable of primary-line and advanced services are
considerable. Key to VoIP primary-line services are robust
quality of service (QoS) standards. The inclusion of
proprietary QoS features in cable modems coming onto the
market probably means that standards will happen sooner
rather than later. CableLabs' approach is to avoid reinventing
the wheel. It tweaks and standardizes the best proprietary
system available and develops technology only if nothing exists
in the marketplace.

Typically, CableLabs creates specifications by combining and
tweaking what it considers the best elements of techniques
already in the marketplace. It will not lack for candidates on
the QoS front. Phasecom Inc. (Cupertino, Calif.) recently
added bandwidth allocation to its SpeedDemon modems,
Motorola Multimedia Group (Arlington Heights, Ill.) says it
has gear with latencies below 40 milliseconds, and Hybrid
offers a channel-sharing algorithm in its modems. At least one
vendor--Com21 Inc. (Milpitas, Calif.)--has QoS functions
because it uses ATM transport.


The organization also is looking into arranging a marriage
between Internet QoS standards such as resource reservation
protocol (RSVP) and cable QoS. "The completion of MCNS
standards with QoS extensions enables cable operators for the
first time to begin exploring deployment of HFC [hybrid
fiber/coax] local loop bypass solutions," says Paul Bosco,
general manager of the cable products and solutions group of
Cisco Systems Inc. (San Jose, Calif.). Guaranteeing QoS may
be the most important task the industry faces. The solutions
that vendors are bringing to market, coupled with CableLabs'
efforts, suggest that the answers could come relatively quickly.
"For the last two months, CableLabs has had a fast-track
effort to get QoS injected into the standard," says David
Waks, president of System Dynamics Inc. (Morris Plains,
N.J.), a firm that consults with cable operators on advanced
services. "That shows how critical this issue is seen."

There are other issues, however, including packetizing dual
tone multifrequency (DTMF) push-button tones and
developing protocols for call initiation, call setup, and the
extension of PBX and Centrex features. Management and
billing issues must be confronted. Gatekeepers--software
devices akin to advanced intelligent network service control
points that act as the link to the H.323 multimedia protocol
and house call control features--must be perfected.

PacketCable is working on a device called a multimedia
terminal adapter (MTA). In twisted-pair VoIP networks, calls
generally enter and leave the customer premises in analog
format. Digitization, compression, and packetization are done
in the network, generally at the gateway device. In cable's
HFC world, the IP network extends to the premises. This
means that the digitizing, compressing, and packetizing of
voice information must be handled by equipment housed at the
end points.

In some scenarios, the PC can provide some of the
functionality--such as its sound card--but this generally adds
too much delay for voice traffic. In the main, the job will be
done by the MTA. These devices will serve as the interface
between the telephone (via an RJ-11 jack) and the Ethernet
system trafficking the data in the home (via an RJ-45 jack).
They will contain an IP stack, a processor, memory, echo
cancellation, and digital signal processors. They will also have
a toggle that will automatically shift operations to the public
switched telephone network if the HFC network goes down,
Craddock says.

The industry must figure out how to deploy MTA circuitry to
the widest group of potential customers. A standalone
MTA--with its own power supply and housing--may be too
expensive. If so, VoIP may be limited to households in which
other services, such as compressed digital video or high-speed
data, allow deployment of a device in which common functions
are shared. Such a scenario would consign the potential VoIP
market to a subset of a subset--targets would be cable
subscribers (it is unlikely that many people who are not cable
subscribers would use cable-based VoIP) who take the other
service or services with which MTA circuitry could share
some processing and powering capabilities. This does not
seem like the best way to get into a new business to many
cable executives. Instead, MTAs costing $200 or less must be
developed, says Jeff Turner, associate director of strategy for
MediaOne (Englewood, Colo.), a cable company.

POWER STRUGGLE

The other vexing issue is power, which is the key to lifeline
service. Today's circuit-switched telephones are powered by
the networks, so they work even if the home's power goes
down. Cable operators must match this to be considered a
primary-line alternative. Right now, only about 4 watts of
power can be sent down the aluminum member of the coaxial
cable entering subscriber homes or businesses. Primary-line
telephony and an OpenCable box would draw 8 to 12 watts.
Since second-line services don't have always-available
mandates, they can be home powered. Finding ways to drive
consumer premises equipment through coaxial cables is a
challenge the industry must face.

If that challenge can be met, the possibility of end-to-end
cable-anchored IP long-distance networks becomes feasible.
In such a scenario, cable operators would be able to offer a
unique set of services. A possible scenario is that calls
between subscribers on the same cable system are free.
Long-distance calls to people who are not subscribers or
aren't served by operators belonging to the consortium (or
another VoIP long-distance consortium with which it peers) go
through a gateway to the public circuit-switched network.
Finally, calls from one operator's participating subscriber to
another would stay within the network. "There's a lot of
interest," Turner says. "It's the Holy Grail for us because you
avoid access fees on both ends."

Cable companies are exploring partnerships with IP voice
startups as well. Tom Evslin, chairman and CEO of IP
telephony provider ITXC Corp. (North Brunswick, N.J.),
says several cable operators have approached his company to
ask about possible deals. "I wouldn't be surprised if there
weren't trials in 1998, but I would be extraordinarily surprised
if there were any significant revenue impact," says Evslin, who
won't divulge the cable companies that have approached
ITXC. "I anticipate commercial service in 1999." Amy Reiber,
the manager of product marketing for The Williams
Companies (Tulsa, Okla.), says teaming with cable operators
on end-to-end IP long-distance networks is an interesting idea
but that no substantive conversations are ongoing.
Some
operators are interested in letting the long-distance carrier
handle per-minute billing issues, which have always stumped
cable operators, Evslin says. Local services, he says, will likely
be either free or charged at a flat rate.

Cable providers have always seen telephony as a natural
evolution of their business. It has been retarded by the
industry's inertia, as well as tricky technical and thorny
regulatory issues surrounding primary-line telephony. VoIP
may well be a way around the biggest problems. Indeed, there
appears to be no single issue that will keep operators from
harnessing VoIP. "I don't see any showstopper," says
Craddock. "I don't think we're going to fall on our swords. It's
just a matter of making it work."

All use of this service is subject to the Terms and Conditions of Use.
All Rights Reserved.

Copyright c 1998 tele.com, The McGraw-Hill
Companies, Inc.
All Rights Reserved.
Website designed by COMPUGRAPHIA
Home page designed by Dennis Ahlgrim.
Last Modified: 5-May-98



To: Greg Jung who wrote (46861)5/14/1998 3:00:00 AM
From: djane  Respond to of 61433
 
5/98 tele.com. [ATM switching as 50%/90% of public network by 2005/2010]

teledotcom.com

CONVERGENCE COUNCIL

The big question right now for public network service
providers isn't if, but when: When will the circuit-switched public network be superseded by a packet-based
infrastructure? Advocates of an IP-centric world say the
change can't happen fast enough, although only the staunchest
'Netheads will go so far as to say that the transformation will
occur in the next couple of years. Most established public
network providers--a.k.a. the Bellheads--are now making
plans to phase out tried-and-true circuit switching for a
packet-based architecture, but they see plenty of life left in
their legacy model. And, of course, the 'Netheads and the
Bellheads are just in the early stages of what promises to be
the grand debate that will shape the destiny of the public
network into the next century: Will IP or ATM (asynchronous transfer mode) be the technology of choice for the
packet-based public network? The "IP on everything" crowd
says ATM will be superfluous once the technology is in place to guarantee IP quality of service; ATM advocates say they
aren't convinced that IP QoS will ever be able to deliver the
reliability required for the public network.

* "Within 10 years, we'll see more than 50 percent of
public-switched traffic carried by packet," says Dan Spears,
the research director at BellSouth Science and Technology
(Atlanta). "Five years from now, we'll see deployments of
ATM packet tandem switches that will be handling voice
traffic, and in 10 years, these switches will be widespread
enough that the majority of traffic will be packet-based."

However, Spears says we should still expect to see circuit
switching being used to some degree 10 years from now.

* Lawrence Vaston, president of the consultancy Technology
Futures Inc. (Austin, Texas), says the public network will
definitely become predominantly packet-switched, but he
expects the change to be incremental, rather than sudden, and
he's convinced that ATM and not IP will be the technology behind this advancement. "We perform forecasts primarily
regarding ATM switching, and this seems to be the technology
of choice for switching voice as well as data," he says. "As far
as what we see for ATM in the public network, it will be
something like 20 percent by 2002. Perhaps by 2005, we'll
see it at 50 percent. But in terms of ATM being the
predominant technology, say 90 percent of the network, we
won't see that until 2010."


* Expect to see packet switching soon, says Jonathon Plonka,vice president of IP architecture and engineering at Frontier
GlobalCenter (Sunnyvale, Calif.). "In the next six months we'll
see significant movement," he says. In two to three years, most
trunking will be packet-based, but because of current
investments in equipment, the edge will remain
circuit-switched.

All use of this service is subject to the Terms and Conditions of Use.
All Rights Reserved.

Copyright c 1998 tele.com, The McGraw-Hill
Companies, Inc.
All Rights Reserved.
Website designed by COMPUGRAPHIA
Home page designed by Dennis Ahlgrim.
Last Modified: 12-May-98



To: Greg Jung who wrote (46861)5/14/1998 3:06:00 AM
From: djane  Respond to of 61433
 
5/98 tele.com. [Info on VPNs and BPP alliance]

Excerpt: "BPP's founders say their ultimate goal is to deliver uniform QoS across different ISPs. "This is really about whether ISPs can compete with the regional carriers for VPNs and IP telephony," says Mike Gaddis, executive vice president of Savvis. "If we don't take this step to build a quality system, we can't win that competition."

teledotcom.com

Peer Pressure Builds on the Internet

Without uniform exchange methods, smaller
ISPs may not be able to meet demand for
OSS electronic flow-through

By Peter Lambert, Senior Writer

Quid pro quo became more like quid pro no over the past
year, most big national Internet backbone providers decided
to supplant much of their free traffic peering arrangements with
other Internet providers in favor of either private peering
limited to ISPs of their own size or paid transit from lesser
players seeking a path through one national backbone to
others.

The move to private peering has been sudden and swift:
According to executives at three national ISPs, more Internet
traffic is now being exchanged via private peering than via
public peering--a first for the formerly egalitarian 'Net.
Although the big ISPs assert that the shift has raised the quality
of ISP-to-ISP network performance, most agree that peering
requirements remain nonuniform, leaving large and small
providers at odds and Internet quality of service (QoS) still
more of a crapshoot than a sure thing.

Unless that changes, ISPs are going to have trouble cashing in
on growing corporate appetites for virtual private network
(VPN) services, observers say. "If you can't do VPNs, you're
a nonplayer in the Internet," says Tom Nolle, president of the
CIMI Corp. consultancy (Voorhees, N.J.).


The challenge now is to strike a better balance between the
quality of private peering and the openness of public peering.
National backbone provider PSINet Inc. (Herndon, Va.) is
now offering what it calls free peering to its network to smaller
providers. PSINet's aim is to counteract "the potential of other
national ISPs to quietly slow down the traffic of other ISPs
until it drives them out of business," says William Schrader,
PSINet's chairman and CEO. Still, while any small ISP can
access PSINet's backbone for free, that would only cover
what Schrader estimates is one-tenth of the 'Net's users and
content providers. For a small ISP to use PSINet's private
and public exchanges to peer upward to reach major
backbones, they have to pay a hefty transit fee ($2,500 a
month for a T1 line) to PSINet.

As another option, four national service providers have founded the Brokered Private Peering (BPP) alliance. The
four--Exodus Communications Inc. (Santa Clara, Calif.),
Savvis Communications Corp. (St. Louis), Electric Lightwave Inc. (Vancouver, Wash.), and The Williams Companies
(Tulsa, Okla.)--propose setting up a nonprofit organization
that would define true peers.
The definition would be based on
scope (national, regional, or local) and traffic type (access,
dial-up, or Web hosting).

Equal treatment

To qualify at each level, a true peer must offer a certain
amount of capacity in a minimum number of markets and serve
a minimum number of customers. Based on those definitions,
national peers would peer for free with other national peers,
regionals with regionals, and locals with locals. As with
PSINet's deal, peering up a level in the BPP hierarchy would
require ISPs to pay a fee. That fee would be close to cost,
according to Rob Bowman, director of backbone engineering
at Exodus.

BPP would enforce exchange policies among members. For
example, if any two peers were to exceed assigned bandwidth
and threaten to saturate an exchange local loop shared with
other providers, those two peers would have to foot the bill to
expand their own exchange facilities. Under the plan, various
BPP members would manage backbones and up to 12
regional exchange points for all members.

BPP's founders say their ultimate goal is to deliver uniform QoS across different ISPs. "This is really about whether ISPs can compete with the regional carriers for VPNs and IP telephony," says Mike Gaddis, executive vice president of Savvis. "If we don't take this step to build a quality system, we can't win that competition."

More than two dozen national, regional, and local ISPs have
expressed interest in BPP, says Matthew Bross, chief
technology officer for Williams. The original founders were
seeking to close charter membership by April 24, then to
schedule the first, formal organizational meeting for mid-May.

All use of this service is subject to the Terms and Conditions of Use.
All Rights Reserved.

Copyright c 1998 tele.com, The McGraw-Hill
Companies, Inc.
All Rights Reserved.
Website designed by COMPUGRAPHIA
Home page designed by Dennis Ahlgrim.
Last Modified: 5-May-98