Blurring The Lines Are CTI, DLC Markets on a Collision Course? [ASND references]
soundingboardmag.com
By Gary Kim
Major changes in networking architecture create whole
new industries.
The $380 billion market capitalization for just 10
networking firms is one example. So is a worldwide
networking equipment market that will hit $32 billion
this year, according to International Data Corp.
(www.idc.com). U.S. data transport revenues,
meanwhile, will hit nearly $20 billion by the end of the
year, according to The Yankee Group
(www.yankeegroup.com). All of that, and more,
resulted from a single trend: the shift from mainframe to
client-server computing. That, in turn, necessitated
interconnection of dispersed local data networks,
widespread deployment of personal computers and
databases running over those networks.
We should anticipate similar changes as the developing
"new public network" [ASND term] supplants the older public
network. For starters, we're looking at a network of
networks, not a single entity. Internet service providers
(ISPs), competitive local exchange carriers (CLECs),
wireless, long distance providers, private networks of
all sorts and services providers who operate no
networks of their own are part of the fabric.
And that means new asset value creation. Already, the
market capitalization for just three new long distance
firms tops $19.5 billion. The market value of four
personal communications services (PCS) providers
tops $4 billion, while 13 public U.S. CLECs have a
market worth of $26.75 billion.
Shifting Architectures
Even within discrete networks, architectures are
shifting. It used to be that distinct network
functions--transport, switching, access--were handled
by distinct parts of the carrier organization, while
vendors maintained distinct development and sales
organizations to match. No longer. The line between
switching and access, for example, has begun to blur.
Increasingly, switching functions are incorporated into
access terminals. Routers and edge switches are
examples. Routing switches and switching routers are
other examples. Internet protocol (IP) software offers
a further example of addressing (switching) functions
moving to the customer premises equipment.
U.S. Internet Appliance Shipments (in thousands of units)
Device
1997
1998
1999
2000
2001
2002
%growth
PC
31,478
36,323
41,576
46,397
51,139
55,590
12.2
Others
1,433
3,634
8,330
16,589
26,439
41,786
96.3
Total
32,911
39,957
49,916
62,986
77,578
97,776
24.3
Source: Dataquest
Even traffic aggregation devices, which remove
Internet traffic before it passes through a standard end
office switch, are examples of this trend, since they
simply attempt to comply with switching decisions
made at the network edge, and pass the traffic along to
the backbone transport layer without further
processing.
Industry merger and consolidation moves are further
evidence, as firms such as Northern Telecom Ltd.
(Nortel) (www.nortel.com) snap up companies such
as Bay Networks Inc. (www.baynetworks.com),
while Lucent Technologies Inc. (www.lucent.com)
acquires suppliers such as Yurie Systems Inc.
(www.yurie.com). The Siemens (www.siemens.com)
linkages with Newbridge Networks Corp.
(www.newbridge.com), and Ericsson's
(www.ericsson.com) anticipated acquisition of data
networking assets are other examples.
The move to "IP over SONET" (synchronous optical
networking) technology also is part of this shift. When
traffic moves directly from a local area network (LAN)
router to the SONET backbone, with no encapsulation
into asynchronous transfer mode (ATM) format, and is
recovered at the destination LAN by a similar IP
router device, switching functions have moved directly
to the premises and network edge.
Another background trend occurring is in instances
where the older circuit-switched network tightly
couples applications and network transport, the new
packet switched network easily separates--indeed,
encapsulates and decouples--network transport from
switching and end-user functions.
In other words, where in the older public network
applications sit on a closed end-office switch, in the
new public network applications can reside anywhere
it is convenient and cost-effective. That might be on a
user's personal computer (PC) or other device, a LAN
server, backbone network server in a gateway, router
or edge switch.
This is no subtle shift. It's a fundamental
transformation, because application development no
longer is controlled by network suppliers or carriers.
Instead, creativity shifts decisively to end users.
Simply, open networks mean faster innovation and
lower cost.
So one must wonder what other architectural shifts
may occur, especially as IP voice becomes a larger
portion of overall networking traffic. Indeed, the
migration of switching functions to the network edge
can be seen in PC proliferation itself, especially as PCs
become communications terminals, launching and
receiving packets with no active intervention by the
circuit-switching network fabric. During the first
quarter of 1998 alone, some 7.96 million PCs were
shipped to U.S. customers, according to IDC.
Public U.S. CLEC Market Value
CLEC
Market Cap., $
Billion
GST Telecommunications Inc.
0.59
Electric Lightwave Inc.
0.94
e.spire Communications Inc.
1.15
ICG Communications Inc.
1.36
Intermedia Communications Inc.
1.61
McLeod USA Inc.
3.32
NextLink Communications Inc.
1.73
RCN Corp.
0.73
Teleport Communications Group
Inc.(AT&T Corp.)
10.84
Teligent Inc.
1.68
US Networks
0.31
WinStar Communications Inc.
1.80
Total
26.06
Source: Goldman, Sachs & Co.
LAN gateways, IP telephony gateways, routers and
edge switches are other examples of edge switching.
Computer-telephony integration (CTI) is another
example. PCs can function as voice terminals, while
the "un-PBX," a workstation-based private branch
exchange, is coming. In fact, the Internet itself is driving
the development of a dizzying array of new devices at
the network edge, including digital cable TV set-top
boxes, web-enabled telephones and personal digital
assistants and video game consoles. By 2002, nearly
half of all terminals will be "non-PC" devices, predict
analysts at Dataquest (www.dataquest.com), a unit of
Gartner Group Inc.
In a real sense, PCs and web-enabled devices
themselves become signal switching devices. And PCs
and other web devices connected to LANs may find
themselves communicating with premises-based traffic
aggregation devices that combine CTI and digital loop
carrier (DLC) features. IP gateways supplied by firms
such as Cisco Systems Inc. (www.cisco.com), Ascend
Communications Inc., Bay Networks Inc. (Nortel),
NetSpeak Corp. (www.netspeak.com), VocalTec
Communications Ltd. (www.vocaltec.com) and
Inter-Tel Inc. (www.inter-tel.com) are examples, says
Raymond James Financial (www.rjf.com) analyst Phil
Leigh.
So as the convergence of access and switching
continues, one might ask whether DLC technology,
now an outside plant access technology, and
burgeoning computer telephony capabilities at the
customer premises, more traditionally the realm of
applications developers, might start to overlap.
"At a bare minimum, functions are being integrated,"
says Josh Soske, CEO of Vina Technologies
(www.vina-tech.com). "Historically, the key system
and the DLC were the two key devices. As T1 bypass
develops, the PBX is turning into the un-PBX."
The demarcation between network and customer
premises is "an intelligent device with an open
interface," adds Tom Barsi, Vina vice president.
Indeed, Vina "pitches itself as a low-end DLC," Soske
says . "We convert analog to digital and aggregate
lines."
Granted, executives in the loop carrier and CTI market
spaces do not today see themselves as being in the
same business. But neither did Lucent and Cisco.
So to the extent that IP telephony is more than public
network switching, and fundamentally embraces
applications, we should see more blurring of lines
between functions embedded in DLC terminals and
those typical of premises-based servers.
Before that happens, each industry segment would
need to redefine its mission. Today, DLC is used to
aggregate analog local phone lines and haul them from
neighborhoods back to central office switches.
Access, rather than call processing, is the key function,
and the leading suppliers are all from the telephone
industry. Lucent Technologies, DSC Communications
Corp. (www.dsccc.com) (recently purchased by
Alcatel), Nortel, Advanced Fibre Communications
Inc. (AFC) (www.fibre.com), Reltec Corp.
(www.relteccorp.com) and E/O Networks
(www.eonetworks.com) are suppliers of DLC
technology.
CTI, on the other hand, tends to be used by suppliers
of key systems and PBXs, call center technology
suppliers and other firms that supply enhanced calling
features including:
Database interaction applications
Audiotex
Fax-on-demand
Interactive voice response
Interactive fax
Fax and voice messaging
Unified messaging, which presents e-mail, fax
and voice messages through one screen
interface and converts from e-mail to fax or
voice for remote telephone access
Intelligent call control for call centers, help desk
automation
Conferencing
Call screening and forwarding
Voice over the Internet
Fax over the Internet
Selected Networking Company Market Capitalization, $
Billion
3Com Corp.
12.54
Ascend Communications Inc.
8.84
Bay Networks Inc.
5.28
Cabletron Systems Inc.
2.10
Cisco Systems Inc.
77.83
Fore Systems Inc.
2.32
Intel Corp.
131.96
Microsoft Corp.
109.13
Oracle Corp.
25.46
Newbridge Networks Corp.
5.14
Total market capitalization
380.60
Source: Company reports, Everen Securities
estimates
We commonly think of DLC as an outside plant,
network access technology and CTI as a
premises-based applications platform. But Dialogic
Corp. (www.dialogic.com), for example, also has
shipped its technology into 80 public networks,
representing more than a million ports, says Jim Machi,
Dialogic's IP telephony director. And especially as IP
communications become more important, one has to
wonder whether the analog-digital signal conversion
routinely supplied by both DLC and CTI suppliers
conceivably could lead to new levels of functional
convergence. To put it another way, could the two
distinct sectors find themselves in the same business, at
least in part?
"Absolutely," says Jay Shuler, AFC's marketing
director. "We are pushing the point at which voice is
converted back to analog form closer and closer to the
customer."
IP telephony gateway products sold by suppliers
including Lucent, Nortel and Vienna Systems Corp.
(www.viennasys.com) are examples, he says.
But despite the blurring of lines between premises and
public network traffic switching and access, many do
not see a conflict looming between DLC and CTI
suppliers.
"CTI companies like Natural Microsystems and
Dialogic will continue to be suppliers to the Lucents of
the world," Leigh says.
Steve Frankel, analyst with Adams Harkness & Hill
Inc. (www.ahh.com), agrees. Companies such as
Dialogic and Natural MicroSystems Corp.
(www.nmss.com) are in the component business, not
the systems business.
"They're the interface between the analog world and
the computer," he says.
In the same way, DLC is an interface between the
analog telephone and the digital transmission network.
And the big telecom suppliers aren't about to let their
business be attacked by upstarts, says Bill Benson,
analyst with Benson, William Blair and Co. LLC
(www.wmblair.com).
"The big players are scary, and won't let their business
go away," he says. "They'll take out the smaller players
rather than lose to them."
So, to the extent that CTI capabilities become strategic
for key suppliers of public networking technology, the
CTI firms will simply be acquired by the telecom
giants.
New Carrier Market Capitalization, $ Billion
IXC Communications Inc.
2.36
Qwest Communications Inc.
8.36
Level 3 Communications Inc.
8.79
Subtotal, IXCs
19.51
Omnipoint Corp.
1.56
Western Wireless Corp.
1.59
Aerial Communications Inc.
0.45
Powertel Inc.
0.80
Subtotal, PCS
4.40
Total, PCS and IXC
23.91
Source: Janco Partners
Big and Bigger
Any way you look at it, the access and CTI market
spaces are big, and getting bigger. In 1998, U.S.
carriers will install some 13.8 million new access lines,
says AFC's Shuler. And a healthy chunk of those lines
will use DLC technology.
At the same time, "upwards of 5 million CTI ports will
be shipped into the enterprise market," Dialogic's
Machi says. All of which makes CTI about a "$6
billion to $8 billion annual market globally," says
Howard Smith, analyst with First Analysis Corp.
(www.analysis.com).
And, one way or the other, each segment has a role to
play in supporting "new forms of access," Machi says.
"The DLC suppliers just have to figure it out." In
addition to supporting ATM, frame relay, digital
subscriber line (DSL), SONET and IP traffic, DLC
suppliers may find themselves "pushing the edge of the
packet network closer to the consumer," Shuler says.
And that's where they might run into CTI providers
taking analog voice and converting it straight to IP
traffic.
"You'll have convergence to the desktop over the long
run, where there's no separate signaling system 7
(SS7) or advanced intelligent network infrastructure,"
says Charles Robins, Pennsylvania Merchant Group
Ltd. managing director. "Longer term, as alternate
networks are built, you won't need gateways. We'll
have DSPs at desktops then, allowing switching right
at the PC." Not to mention PBX functionality that runs
across a company intranet.
The cost of bandwidth and processing power are the
two fundamental engineering realities dictating the pace
at which applications and functions can be migrated
from one place in the network to another. To put
matters another way, as chip costs drop, and local
bandwidth increases, the convergence of DLC and
CTI is more likely.
"Economics now dictate that no expensive digital signal
processors (DSPs) can be put into the customer
premises equipment," Shuler [AFCI] says.
Computing horsepower, for example, conditions a
PC's ability to function effectively as a voice terminal.
"Little things, like the operating system and the chip
you use for digital-to-analog conversion and
packetization, affect voice quality," Shuler says. In
general, consumers will experience higher quality when
a dedicated packet conversion device is used, he
notes. That's because the packet voice system doesn't
have to contend for the host PC's processor and
memory.
Access pipe characteristics make a difference as well.
More bandwidth, or the ability to control bandwidth, is
the key. And that suggests ATM and IP
quality-of-service (QoS) mechanisms will help push
convergence of the CTI and DLC functions. Since
voice quality is dramatically affected when the packets
have to contend for access with huge data packets,
some means of controlling access is really helpful.
ATM is one way to do so. So are various QoS
mechanisms suggested for IP traffic. As a rule, as long
as bandwidth remains limited, QoS assurances
provided by ATM are more important. As bandwidth
grows, IP works increasingly well.
Though many would disagree, the CTI and DLC
industries "are on a collision course," Shuler says.
"Two things are happening," says Herb Tinger, First
Albany Corp. (www.fac.com) senior vice president.
"You have the convergence of voice and data, plus the
opening of traditional proprietary telecom
infrastructure."
And that means "thousands of software developers"
ultimately will be unleashed to write applications
spanning the access and applications markets. "The
real value of a company like Natural MicroSystems (a
supplier of DSP hardware and software) is the
software architecture," says Ted Jackson, Piper Jaffrey
analyst. "It's ideal for third-party application
developers." And that means a "company like Lucent
gets hardware and software tools to build applications
fast," Jackson says.
All of which will make easier the task of bundling
additional functions into DLC terminals, as well as
adding "voice services to customer premises
equipment," says Karen Barton, vice president of
Xedia Corp. (www.xedia.com). That could spawn a
whole new equipment market space.
Gary Kim (garykim@concentric.net) is a
contributor to Sounding Board magazine.
Copyright c 1998 by Virgo Publishing, Inc.
|
