Will Cisco take market share from circuit switch vendors such as Lucent and Nortel? Here is a contrarian view. The following excerpt from "IP Myth vs. Reality" argues most switches purchased in the next five years will be from Lucent or Nortel.
" Router vendors will win over conventional circuit switch vendors regarding 21st Century IP
telephony products.
Router vendors such as Cisco dominate IP switching for information services. Alternatively,
dominant circuit switch vendors such as Lucent and Nortel don't have roots in IP, although they
are trying to play catch-up by buying small router and/or IP vendors. The strategic question here
is: Will it be easier for router vendors to solve the transition from circuit to IP switching for 21st
Century carriers than it will be for a circuit-switch based vendor?
I'll place my bet on the circuit switch vendors with regard to public telephony. Here's why:
Circuit-switch vendors (Lucent, Nortel, etc.) have invested hundreds of millions of man-hours and
dollars creating switch interface software products in the billing, customer care and OSS areas for
the North American carrier marketplace. If foreign global circuit switch giants like Alcatel, Siemens
and Ericsson have failed to make significant inroads in Lucent's and Nortel's home turf, and vice
versa for Lucent and Nortel in Western Europe, then why are router vendors so optimistic that they
are going to eat the lunch of these giant circuit-switch vendors with a new switching concept (IP)
and no OSS experience? My prediction is that the vast majority of new public or carrier network
switches purchased for the North American market in the next five years will have Lucent or Nortel
logos. This prediction doesn't necessarily apply to access switches needed to process IP traffic
between xDSL and wireless and/or cable modem end users accessing public circuit-switched
networks.
On the other hand, there are the ATM backbone networks front-ended with ATM edge switches
(hybrid router/ATM switches) used to support IP backbone networks and private ATM networks.
Will router vendors such as Cisco dominate edge switching for private networking and Internet
backbones for the next five years? Probably yes! Just as it was a myth to think computers and
telecommunications were merging in the 70s and 80s, and that IBM would eat AT&T's marketplace
lunch, or vice versa, it's a myth to think that router vendors will dominate the circuit switch vendors
or that circuit switch vendors will gobble up the router marketplace anytime soon."
IP Myth vs. Reality - Part 2
Posted November 03, 1998 10:00 AM PST
This article is the second of two on this subject. Check the Carriers tab on your Telecom Insider for
IP Myth vs. Reality - Part 1. Here, Dr. Jerry Lucas of TeleStrategies, Inc. reviews what is fact and
what is fiction in the world of IP telephony.
IP telephony carriers' greatest challenge is interconnection with circuit switched carriers.
On one hand, IP telephony equipment and carrier interconnection will be a challenge. For example,
interfacing traditional carrier SS7 networks with IP gateway switches won't be a trivial task.
However, the IP carriers' greatest challenge will be interconnecting with other IP carriers.
The difficulty lies within the traditional IP backbone provider peering arrangements, and the fact
that the Internet is not regulated.
The FCC cannot force peering relationships among the big three IP backbone carriers (UUNET,
Sprint and MCI-or soon-to-be new owner Cable & Wireless) that control about 80 percent of IP
backbone traffic. The big guys peer among themselves with bill and keep arrangements--you
terminate my customer's traffic on your network and I'll do the same for your customer's traffic on
mine. However, the big IP wholesalers don't peer with the little ISPs. They charge them for what's
called transient traffic, for example, $4000 per month per T-1 to terminate an originating ISP's traffic
on another backbone network. Note that MCI had to spill the beans on its network infrastructure
and number of customers when they put their IP backbone network up for sale.
MCI identified 40 peering arrangements and 1300 ISP transit arrangements. The bottom line, this
peering practice was OK when the Internet was used primarily for information access and e-mail,
because the inter-backbone traffic was roughly symmetrical and "bill and keep" made sense. In the
new IP world of multimedia, traffic will be highly asymmetrical, so that bill and keep will not work.
An accounting and billing system based on usage is needed among IP carriers for many reasons.
Replacing the system of peering and settlements will require IP carriers' cooperation and more, and
it's not going to happen overnight.
AT&T is introducing IP telephony, so it must be the way to go.
AT&T introduced its domestic Connect-n-Save this May with calling card-like features and
discount usage rates (7.5 cents per minute). So they must believe IP telephony is cheaper, and that
IP, technologically, is the way to go. I don't think so.
My feeling is that this IP telephony offering is nothing more than a signal to the FCC. In other
words, AT&T is saying, "If you, FCC, don't regulate these small, start-up IP Telephony carriers and
make them pay access charges (4 to 5 cents per call), we will screw up the works regarding the
Universal Service Fund and go the unregulated, IP telephony access route ourselves."
Router vendors will win over conventional circuit switch vendors regarding 21st Century IP
telephony products.
Router vendors such as Cisco dominate IP switching for information services. Alternatively,
dominant circuit switch vendors such as Lucent and Nortel don't have roots in IP, although they
are trying to play catch-up by buying small router and/or IP vendors. The strategic question here
is: Will it be easier for router vendors to solve the transition from circuit to IP switching for 21st
Century carriers than it will be for a circuit-switch based vendor?
I'll place my bet on the circuit switch vendors with regard to public telephony. Here's why:
Circuit-switch vendors (Lucent, Nortel, etc.) have invested hundreds of millions of man-hours and
dollars creating switch interface software products in the billing, customer care and OSS areas for
the North American carrier marketplace. If foreign global circuit switch giants like Alcatel, Siemens
and Ericsson have failed to make significant inroads in Lucent's and Nortel's home turf, and vice
versa for Lucent and Nortel in Western Europe, then why are router vendors so optimistic that they
are going to eat the lunch of these giant circuit-switch vendors with a new switching concept (IP)
and no OSS experience? My prediction is that the vast majority of new public or carrier network
switches purchased for the North American market in the next five years will have Lucent or Nortel
logos. This prediction doesn't necessarily apply to access switches needed to process IP traffic
between xDSL and wireless and/or cable modem end users accessing public circuit-switched
networks.
On the other hand, there are the ATM backbone networks front-ended with ATM edge switches
(hybrid router/ATM switches) used to support IP backbone networks and private ATM networks.
Will router vendors such as Cisco dominate edge switching for private networking and Internet
backbones for the next five years? Probably yes! Just as it was a myth to think computers and
telecommunications were merging in the 70s and 80s, and that IBM would eat AT&T's marketplace
lunch, or vice versa, it's a myth to think that router vendors will dominate the circuit switch vendors
or that circuit switch vendors will gobble up the router marketplace anytime soon.
The new upstart national IP carriers will eat AT&T's, MCI's and Sprint's lunches.
Carriers such as Qwest, Level 3 and Williams have gotten a lot of publicity about the creative ways
in which they have financed their fiber build-outs and their strategies of using an IP backbone for
public telephony and private networking. The myth is that they have networks so fundamentally
different from the incumbents, and no customer base to cannibalize, that they will dominate the
marketplace. The reality is that there is no fundamental difference in backbone networks, new or
established. All have fiber, WDM, ATM backbone, edge switches, etc. In the future, a different
network architecture could arise, but it wouldn't contain any exclusive, secret network
infrastructure silver bullets unknown to other carriers.
Also, I believe that anything the upstart carriers can do in the backbone services and/or pricing
arena, the incumbents can do as well. The only big margins left in the long distance business are
international and local access. The upstarts are just getting their feet wet in these markets, and they
have a long way to go.
For the upstarts to eat AT&T's, MCI/WorldCom's and Sprint's lunches they must build a CLEC
business base. The long distance carrier that has fiber to commercial office buildings will win the
big corporate and private network user business. Note: WorldCom's merger with MCI will further
boost its CLEC fiber penetration, already at 27,000 domestic and 4,000 European buildings.
WorldCom has a transatlantic and Pan-European fiber network to boot. Creating a national fiber
network is only a start.
IP networking is the future and cannot be made obsolete.
Technologies that are obsolescence-proof have come along in the telecommunications industry
every five years or so since competition has been allowed. In the 1980s, IBM, COMSAT and Aetna
poured billions into a venture called Satellite Business Systems (SBS), relying on small,
customer-premise satellite dishes that were going to provide the corporate networks of the future.
In the late 80s, the RBOCs also had an obsolescence-proof technology for which they spent $20
billion in switching software called ISDN. In the early 90s, the telcos again discovered a new
obsolescence-proof technology called ATM. None of these obsolescence-proof technologies came
close to living up to their hype. Today we have IP technology for voice, data and video services.
Yes, IP is in a class by itself vs. SBS, ISDN and ATM, but it's a myth to think that nothing is on the
horizon to make IP obsolete.
For starters, there's dense wavelength division multiplexing. Utilizing this fiber technology was not
practical from an engineering economics perspective even two years ago. Today it stands on the
verge of replacing the vision of ATM and SONET in the backbone for the 21st Century. Consider
that if the only thing ATM technology does today is interconnect 20 or so high-capacity ATM
backbone switches so routers or edge routers can interconnect, why not just connect the fiber
backbones to routers in a ring architecture. Then assign each router an optical wavelength instead
of an ATM packet address. Furthermore, why use SONET to channelize a 2.5 GBPS traffic stream?
Just let the router use the entire 2.5 GBPS traffic stream and identify user traffic by an IP address
instead of position in time, as with SONET.
Go even further: if billions of different wavelengths can be created by different quantum state
transitions in our universe of matter, just assign each person in the world a unique wavelength
instead of a telephone number or IP address. Then switches would be just a big "conceptual prism"
of glass. Obviously there's some work to be done on the quantum physics and laser technology
part, so IP should reign for some time as the leading obsolescence-proof technology candidate.
__________________________
Excerpted from the Publisher's Letter: IP Myth vs. Reality, by Dr. Jerry Lucas, from Billing World
magazine's October 1998 issue. To learn more about billing, customer care, and OSS, visit Billing
World magazine on the Web by clicking on their logo at the top of this article. |
