[Continued. A Sidebar from the preceding article.]
Critical Mass
Now that competition is real, pressure is
building on high-speed ISPs to prove they can
maintain service over the long run.
By Carl Weinschenk cweinsch@cmp.com.
teledotcom.com
-------
Talking time is done. After years of promise, robust
competition is finally reaching the consumer and small-business
markets for high-speed Internet access, thanks to the
continued growth in cable modem services and the emergence
of digital subscriber line (DSL) service. This competition is
expected to blossom in the next several months, with analysts
estimating that the number of cable modems in service will
climb to one million and that the number of installed DSL lines
will track sharply upward from their first-quarter level of
74,000 to 248,000 by year's end.
There's no arguing against the benefits of competition. But it
doesn't mean that with their services finally established, all
these providers need do is sit back and watch subscribers and
revenue fall into their laps. Quite the contrary. This critical
mass is transforming the issues many of these providers face,
forcing them to place new emphasis on challenges ranging
from developing adequate network infrastructure to network
management, marketing and dealing with the outcome of
regulatory changes. How they meet these challenges is no
small matter. In fact, the often markedly different approaches
taken by cable modem operators and DSL providers is likely
to have long-term impacts on the deployment and growth of
both services.
Nowhere is this difference more obvious than in network
access. The cable industry did receive an unexpected shock
this month when a federal court ruled that municipalities are
within their rights to demand open access as a condition of
cable franchise transfers. It's unclear whether the ruling will
affect the industry's long-term investment, but these
competitive services still appear to be heading in different
directions on access. Consider Telocity Inc. (San Jose, Calif.),
a next-generation Internet service provider (ISP) that plans to
roll out an assortment of customer-provisioned Internet
access, voice-over-Internet-protocol (VoIP) and wireless
home networking services beginning this month. The company
will use DSL, but it would also love the chance to gain access
to cable networks as well.
That's not likely, though, says Thomas Obenhuber, vice
president of product development at Telocity. Cable
operators, he notes, are just too reluctant to let independent
ISPs ride their networks. For all the good the cable industry
has done in developing its high-speed data services,
Obenhuber believes this possessive nature could backfire by
ultimately stifling innovation by limiting the number of providers
offering service. All this comes in stark contrast to DSL
service, where new providers are emerging constantly.
"Whenever you have competition, the final product is a better
product," Obenhuber says. "People have to be more creative;
they don't just occupy the consumer."
The cable industry, of course, has its own take on what's good
for competition, and it points as proof to ServiceCo LLC
(Reston, Va.) and its Road Runner service, one of the two
major cable ISPs. The company is throwing time and money
at developing a customer-responsive system that allows
subscribers of its cable-system partners to self-provision
service startup, make real-time changes and be alerted to
discounts on bundled offerings, among other things, says
senior vice president of operations Stephen Van Beaver. He
hopes the effort someday will move to industry research and
development (R&D) consortium Cable Television
Laboratories Inc. (CableLabs, Louisville, Colo.) so that it and
systems under development elsewhere can be standardized
industrywide.
It shouldn't be hard, according to Van Beaver. All the
high-speed data-over-cable companies and their vendors,
which are developing similar systems, are working with
industry specifications and standard application program
interfaces (APIs) to the relatively few operations support
systems (OSSs) and billing systems that serve the industry.
The approach to network access is only one difference and
one challenge. There are others, but they all relate to three
common aims that cable modem operators and DSL service
providers share. They include meeting the market's current
pent-up demand for high-speed Internet access, getting
positioned to fight for additional market share once this
"low-hanging fruit" is picked off and absorbing all this demand
without breaking down network service. This threefold
challenge will create quite a juggling act for providers.
"Both the telephone and cable industries will be faced with
user demand that is far beyond what the original networks
were designed for," says a service provider executive who
requested anonymity. "These are applications that demand
instantaneous, very broad bandwidth bursts of information in
both directions. When you start looking at duplex loading of
megabit and multimegabit services, both networks are going to
come screaming to their knees if they are not very careful in
how they deploy."
A recent study by the consultancy Ernst & Young LLP (New
York) supports the fears of the anonymous executive. It
predicts that local bandwidth demands for eight major
applications, including digital television, e-mail,
videoconference, cable television, and local and long-distance
voice service, will more than double from about 2 million
Tbit/s this year to about 4.5 million Tbit/s in 2003.
This increased demand will be even trickier to manage,
because the traffic landscape will change. Networks will face
the burden of not only more traffic but also a different type of
traffic. The composition will migrate from almost exclusively
error-indulgent Web surfing to traffic increasingly composed of
bandwidth-hogging and/or latency- and jitter-sensitive
applications such as VoIP, instantaneous "twitch" video
gaming, video streaming, live entertainment, business and
educational events, and downloading of trial software, music
and CD-ROMs. This pattern will be sprinkled with traffic from
an assortment of home networking and other applications,
such as home security and automation.
Beyond these changes, the growth of e-commerce and
telecommuting will make security a greater concern. "Doing
high-speed Internet access is a great application, but if you
haven't designed the network and systems to handle other
types of services that are definitely coming down the stream,
you can run into a brick wall," says Rex Cardinale, vice
president of engineering and chief technical officer of Covad
Communications Co. (Santa Clara, Calif.).
Against this backdrop, these competing providers are coming
to realize that most subscribers won't have an inherent
preference for who provides their high-speed access. They'll
simply want it. That leaves providers with the task of setting up
marketing operations plans and back-office processes that are
both efficient and customer- friendly. "Really, it is more of a
fundamental issue of who delivers services and how good a
job of putting the processes and systems to deliver outstanding
customer service," says Jim Anderson, director of product
management for fast access for MindSpring Enterprises Inc.
(Atlanta). The company is in the middle of a five-city rollout of
cable modems with Knology Holdings Inc. (West Point, Ga.)
and will begin offering DSL services in Atlanta with BellSouth
Corp.
To their credit, many providers have been preparing for their
new world. As recently as a year ago, for example, attention
was turned toward the physical infrastructure of cable and
DSL networks. All the while, proponents of either technology
have been flinging shots at the other. To a large extent, these
charges are overstated, and providers appear to be addressing
the legitimate problems.
Challenges to DSL, such as distance limitations, spectral
compatibility and the inherent problems with load coils and
bridge taps, are being addressed. Load coils and bridge taps,
which optimize voice service but can gum up data
transmissions, had been scattered about for so long that
nobody knew where they were. New test gear from Harris
Corp. (Melbourne, Fla.) and others alleviates this
embarrassing problem. Also, some new forms of DSL can
work through these impediments. In addition, competitive local
service providers can get around these problems by simply
specifying that they get loops qualified as being free of such
inconveniences. "Problems seem to have a half-life of no more
than six months each," says Jim Southworth, director of
advanced network services and technologies at ISP
Concentric Network Corp. (Cupertino, Calif.). "I expect that
this interval will remain constant for the next few years."
Likewise, spectral compatibility-the ability of DSL lines to
coexist with 1.544-Mbit/s T1 lines-is another issue that is
being addressed. "Those things are all real, but normal
par-for-the-course kind of things," says Cardinale. "Spectral
compatibility is not new to DSL."
On the cable front, the main question is how quickly the
industry is getting itself positioned for two-way. The signs are
good. It has upgraded about 68 percent of its plant to
two-way operation, virtually all of it hybrid fiber/coax (HFC),
according to Leslie Ellis, an analyst with market watcher Paul
Kagan Associates Inc. (Carmel, Calif.). That number is
projected to rise to 71 percent next year, she says.
Setting up two-way operation is not the final step to
broad-ranging availability of high-speed Internet access
through cable networks. But it is the essential underlying step.
Others include the growing availability of cable modems.
Considering the cable industry has about a 5 percent buy rate
for its Internet access service-and there will soon be a million
subscribers-it's fair to estimate that this cable service is now
available to about 20 million people.
This progress doesn't mean that physical-layer challenges don't
exist. Yet the biggest concerns may not be adressed until the
new era begins: They involve whether the networks hold up to
the increased traffic pressure and, if not, whether providers
will move to make deep changes or try to find quick fixes.
The subtle change in this equation is that there appears to be a
shift away from providers offering a purely technical response
to these challenges and toward those that provide a
business/technology mix. In the cable scenario, many of the
upcoming decisions may be based on divvying up its shared
pipe, which is far from an exact science. For instance, there
may not be enough upstream (home-to-headend) capacity to
support all the applications that customers want. A node may
be working well for surfing, but not for video streaming. In
other cases, everything may be working well until somebody
decides to cybercast a Bruce Springsteen concert. In that
case, what happens to VoIP when "Born to Run" starts?
These will be everyday headaches if networks aren't
engineered correctly. "When you bring on a couple of
streaming videos, it drastically changes the number of
subscribers that can be supported doing other things, like
e-mail or Web surfing," says Terry Wright, chief technology
officer of Convergence.com Corp. (Suwanee, Ga.), a turnkey
data-over-cable company that has agreed to become a wholly
owned subsidiary of vendor C-Cor Electronics Inc. (State
College, Pa.).
A shortage of upstream capacity may require bullet biting,
such as not deploying some services for which there is
demand, instituting capacity limitations or installing
sophisticated and expensive solutions, such as local dense
wavelength-division multiplexing (DWDM), Wright says. The
first two solutions would be just what DSL marketing
departments ordered because it would limit customer offerings
and reinforce the notion that cable service is not user-friendly.
DSL providers have their own challenges, including the similar
issue of enhancing plant facilities to handle new and expanded
demand. For these providers, the bottleneck may be the links
from the central offices (COs) to the core network. The
fundamental question that these providers face is whether
they're willing to ante up to make sure they have enough
capacity. "There are things such as capacity and latency and
delay and lost packets that are strictly in the hands of the
service providers," says Tom Starr, a board member of the
ADSL Forum and a member of the technical staff of
Ameritech Corp. "It's a matter of, Are you willing to provision
your network?"
Despite these questions, it can reasonably be argued that most
consumers will get satisfactory performance from both
networks. Indeed, many of the claims of superiority on both
sides don't withstand scrutiny. A big advantage that Road
Runner and cable's other major ISP, At Home Corp.
(@Home, Redwood City, Calif.), can currently claim is access
to hyper-fast networks based on WDM and rife with caching
and load balancing gear to bring heavily used content closer to
customers. The reality is that these are marvelous networks
and precisely the kind that will be available to DSL through
companies such as Qwest Communications International Inc.
(Denver)-from whom Road Runner is getting its service-Level
3 Communications Inc. (Louisville, Colo.) and the Williams
Network (Tulsa, Okla.). All the ISPs have to do to tap in is
pay up. Likewise, DSL companies say cable data is insecure
because it's a shared media. What they don't say is that their
networks are also shared, just at a point closer to the core.
Ultimately, the competing groups appear to have realized that
most of the charges are overhyped and that what they need to
worry about most is working with the consumer. Covad,
which to date has been a wholesaler of services to business
ISPs, recently announced its first residential offering, which
involves a push in 12 major metropolitan areas. Beyond this,
the carriers recently announced a successful voice-over-DSL
test using the company's asynchronous transfer mode (ATM)
network. The goal is to automate the process as much as
possible, Cardinale says: "If I have to have humans touch all
the orders that pour in, that will slow down the business."
There are signs that these groups are doing more than just
concentrating on customers; they may actually be inching ever
so slightly toward one another. DSL may be the catalyst for
the long-expected shakeout among ISPs, thereby calming a bit
of the fractiousness that the industry often presents as an
advantage. "This is a time when the big service providers and
ISPs will get bigger and the little ISPs who can barely keep up
with dial access will lose rapidly," says Concentric's
Southworth. "DSL will kill the business model of many smaller
ISPs this year."
The cable modem business isn't immune to change, either, as
the recent court ruling indicates and as money starts finding
@Home's and Road Runner's competitors. Two of these
secondary companies got big boosts during the last six
months. In December, High Speed Access Corp. (HSA,
Denver), a turnkey cable modem company, got $20 million
from Microsoft Corp. cofounder Paul Allen and, three months
later, filed a registration statement for an initial public offering
(IPO). Likewise, Convergence.com's deal with C-Cor will
give it more muscle and strong entree with cable operators.
Clearly, neither technology is going away. "I don't think
anyone can rightly deny that when it comes to high-speed
Internet service as well as access to corporate LANs
[local-area networks], both cable modems and DSL-based
modems will command a sizable share of the market," says
Starr. The question that's yet to be answered is, Which one
will dominate? |
