Interesting article on DWDM. On a quick scan I did not see TLAB referenced. How far behind is TLAB to having a product to market that will generate revenues?
Mike
November 08, 1999, Issue: 421
Section: Feature Articles
Surf's Up? -- Momentum Is Building, But When Will Local
Service Providers Catch The Optical Wave?
Jeremiah Caron and Sandra Guy
Optical networking has thrown off a lot of light on Wall Street recently. And
since the mid-'90's, the technology has also generated plenty of heat in the
long-distance market.
There, dense wavelength-division multiplexing dwdm gear is still expensive
due to its complexity. "They're using very sophisticated components," says
Lawrence Gasman of CIR, "and the People making them are Ph.D.-level
engineers."
it's perhaps telling that one of the few pieces of gear currently available,
Cerent's 454, is all about enhancing existing rather than replacing Sonet/SDH
infrastructure.
Optical networking has thrown off a lot of light on Wall Street recently. And
since the mid-'90s, the technology has also generated plenty of heat in the
long-distance market. There, dense wavelength-division multiplexing
(DWDM) technology has worked a multiplier effect on fiber network
capacity to rival the astronomical valuations of optical startups like Sycamore
Networks Inc. (Chelmsford, Mass.). But there's still little heat or light on
Main Street, with hardly any DWDM in local communications networks.
That could soon change. The forces of competition and Internet growth that
have driven down service pricing and driven up traffic on long-distance
networks are working their way into local networks. A growing number of
local service providers are looking to increase their network capacity in
anticipation of a boom in demand. Almost as important, they want to cut
costs and reduce the time it takes to link corporate customers to their fiber
backbones from months to minutes.
That's one reason for all the fuss on Wall Street. At the very least, this new
local mandate means an evolution of Synchronous Optical
Network/Synchronous Digital Hierarchy (Sonet/SDH), which has been the
transmission technology of choice for metropolitan fiber networks since the
1980s. Some say it eventually will mean a complete replacement of
Sonet/SDH networks. There's very little consensus on how to bring local
fiber networks up to par, though, so it may be some time before service
providers can surf optical wavelengths as cost-effectively as they'd like to
right now.
How long is unclear, given the frantic pace of investment, research and
development of optical networking technology. Scores of new and old
companies are mixing various fiber optic transmission and management
concoctions that combine to move glass-and-light-enabled bandwidth closer
to applications and the businesses and consumers that use them. The
emerging solutions generally seek to apply the structure and trustworthiness of
Sonet/SDH with the flexibility of today's switches and routers to the
lightwaves that DWDM produces.
The venture capital market alone has sunk an estimated $7 billion to $10
billion over the past 18 months into the companies trying to speed
deployment. Cisco Systems Inc. (San Jose, Calif.) dropped nearly $7 billion
worth of its stock this past summer on Cerent Corp. (Petaluma, Calif.), a
startup. And when Sycamore's share price drove from $38 to as high as
$270 on the first day of its initial public offering, many Wall Street analysts
proclaimed the beginning of a wave of optical IPOs (see "Sycamore Takes
Root," page 44).
Established equipment suppliers are also putting forth their particular
solutions. Among them are DWDM powerhouses such as Alcatel N.V.,
Ciena Corp. (Linthicum, Md.), Lucent Technologies Inc., NEC (Tokyo),
Nortel Networks Inc. and Fujitsu Network Communications Inc.
(Richardson, Texas).
For their part, local service providers are content to wait and see, with few
agreeing to discuss their future archi- tectures for this article. Still, between
1998 and 2002, growth in the North American optical networking market is
expected to come more from local than from long-distance networks,
according to research firm Ryan Hankin Kent Inc. (South San Francisco,
Calif.). Growth in local spending will rise from zero to more than 11 percent
of total spending on optical networking, while long-distance growth will slow
from 28 percent to just under 4 percent (see "Small Swells"). The total North
American market is expected to reach $3 billion in 2002.
There's a lot of talk in the market about so-called "metro DWDM." This
implies replacing Sonet/SDH add/drop multiplexers (ADMs) with
lighter-weight versions of the long-haul DWDM transmission systems and
then somehow mapping protocols, such as the Internet protocol (IP) and
asynchronous transfer mode (ATM), directly to the wavelengths that such
devices produce.
To many, this all-optical approach represents nirvana, the ultimate destination.
But the reality is that it is a strategy of replacement, not evolution, and a
prohibitively expensive one at that.
Current DWDM configurations like this may not be ready to handle the
diverse requirements of metro-area traffic, either. Local networks have much
shorter distances between nodes and a lot more "hop on, hop off" traffic than
long-distance networks. A home for the all-optical approach does exist
today, but it's in point-to-point long-haul and regional networks, say service
provider technologists. Enron Communications Inc. (Portland, Ore.) is one
national network operator deploying its PureIP fiber optic backbone by
linking switch routers directly to its DWDM transmission systems.
For now, most local service providers aren't buying; in fact, many of the new
optical products intended for the local market aren't yet ready for sale. Still,
many carriers are evaluating this remarkably diverse range of options under
development.
In existing metropolitan fiber optic rings, the spans between nodes are usually
less than 20 kilometers (12.5 miles). These nodes typically consist of
add/drop time-division multiplexers (TDMs), digital cross-connects and
management paraphernalia. DWDM gear, like Sonet/SDH gear, could
possibly be positioned at these nodes, distributing traffic over multiple
wavelengths rather than sending it over a single wavelength in multiple time
slots, as prescribed by today's dominant TDM architectures. This would
increase the existing rings' capacity much the same way that DWDM has
bolstered long-haul segments.
Many assert that metropolitan-area Sonet/SDH rings are filling up with traffic,
making the prospect of fuel-injecting their capacity with DWDM attractive.
But at what cost? It's widely believed that service providers today can more
cost-effectively lay additional fiber, assuming it can be pulled through existing
conduits, than replace or enhance Sonet/SDH ADMs. The average local
provider can pull fiber for $30,000 per mile, according to a BankBoston
Robertson Stephens (San Francisco) study.
DWDM gear is still expensive due to its complexity. "They're using very
sophisticated components," says Lawrence Gasman, president of
Communications Industry Researchers Inc. (CIR, Charlottesville, Va.). "And
the people making them are Ph.D.-level engineers." Also, DWDM is a
point-to-point technology that uses fixed-wavelength lasers and receivers. To
get the flexibility and reliability desired in a metro environment, each node
would need a direct link to every other one-an economic impossibility for a
network of any reasonable size.
Time is an issue. The consensus is that it takes six months or more for a
carrier to build out a ring from its central office (CO) when a customer
demands an OC-12 (622 Mbit/s) or higher-speed interface. Thus the push to
apply internetworking principles to wavelengths, which would allow faster,
centralized provisioning and more service flexibility. Others maintain that the
real problem isn't capacity at all but rather getting customers high-speed
access to the already existing fiber.
In fact, the two considerations-metro backbone and metro access-are
interrelated. Changes in the architecture of one will affect the other. The
more-established equipment suppliers are tackling both areas with product
plans or acquisitions, while the startups are generally split between the two.
The approaches vary widely; these are uncharted seas with no solution
holding a clear advantage. Also, most of the metro-area optical systems are
just now in testing, fresh off the drawing board.
The many approaches to metro-optical networking defy easy categorization.
Still, beyond the basic distinction between solutions for metro backbone and
metro access, it's possible to discern the following groupings: Sonet/SDH
enhancements or improvements; DWDM "lite" and hybrids that mix DWDM
with transport-layer protocols while lowering DWDM gear costs; and "none
of the above"-entirely new approaches.
It's perhaps telling that one of the most prominent exceptions to the
"not-yet-for-sale" rule for metro-optical networking gear, Cerent's 454, is all
about enhancing existing rather than replacing Sonet/SDH infrastructure. The
system aggregates multiservice protocol traffic, including IP, ATM and frame
relay, and a wide variety of service-side optical interfaces, at all bit rates. It
uses a small box that Cerent says saves equipment costs and increases
provisioning flexibility and management capability. |
