Posted: 06/1999

Crossing the Pond
Global Carriers Build, Buy or Lease Undersea Cable
By Ken Branson

The bottom of the North Atlantic has been accumulating submarine cable since 1858, when
the first one was laid between North America and Europe–a telegraph cable. Since then,
coaxial and then fiber optic cable has snaked under the water.

For more than a century, the business of planning, engineering, laying and maintaining
submarine cable proceeded at a stately pace. Huge carriers on either side of the water
formed "clubs"–AT&T Corp., let’s say, on one side and London-based British
Telecom-munications plc on the other–to fund the work. Voice traffic, which is what the
cables were mainly meant to carry, grew at a healthy but predictable rate each year.

The Internet and the demand for data capacity has changed all that. In 1997, when what
was then WorldCom Inc. (now MCI WorldCom Inc.) and Cable & Wireless plc, London,
announced that they would build an undersea synchronous digital hierarchy (SDH) ring under
the Atlantic between New Jersey and Wales, they made it plain that they were doing it
because of data and the Internet.

"Traditionally, we used to plan cables for 90 percent voice and maybe 10 percent
data," says Seth Blumenfeld, MCI WorldCom’s president of international correspondent
and carrier services. "That paradigm has changed dramatically. For large carriers,
it’s 80 percent for wideband data and 20 percent for voice."

Meeting Bandwidth Demand

The technology for providing bandwidth has kept up with the demand for capacity–it
might, in fact, be said to have stimulated the demand, since land-based competitors
already have begun to use it in their terrestrial networks. Dense wavelength-division
multiplexing (DWDM) has made it possible to get more wavelengths, and hence more
information, down an optical fiber. Improvements in zero-dispersion optical fiber have
meant that the fiber itself is more robust.

"The reason DWDM is so significant is that it’s so upgradable," says John
Beagly, vice president for market development in the optical networking group, Lucent
Technologies Inc., Murray Hill, N.J. "The cost of maintaining submarine cable is
high. And the cost of going down there and laying in more cable when you need it … well,
with DWDM, you don’t have to go down there at all."

The upgrades can be accomplished from the beach, and accomplished less frequently. DWDM
systems now being deployed in the oceans have 16 wavelengths per fiber, and some of those
being announced have 32 wavelengths. The more channels one has, of course, the less power
per channel. Lucent and other manufacturers of undersea cable are developing
zero-dispersion fiber that requires less power per channel.

Graph: Cumulative Investment in Submarine Fiber Optic Systems

"The wild thing," says David Martin, vice president-business development, The
CTR Group, Woodcliffe Lake, N.J., "is that, given the huge and unbelievable increase
in capacity of the local cables, they’re still proliferating as fast as they always did. A
year ago, we were advertising that our cable would carry 320 gigabits per second (gbps)
crossing an ocean. Last fall, we announced the state of the art had moved to 640gbps. In
January, it was 1.28 terabits–a fourfold increase in capacity!"

Martin, whose company is working on Project Oxygen, a worldwide undersea cable project,
is a subscriber to what might be called the conduit-of-dreams theory: "As soon as
people get cheap bandwidth in front of them, they’ll find a way to use it."

Indeed, WorldCom’s annual report for 1997 said Internet traffic was growing 1,000
percent per year, compared with 7 percent to 8 percent per year for voice. Data is where
the money is, and where the capacity is used.

Project Oxygen plans a global network–168,000 kilometers in its first phase, of which
153,000 will be undersea. That phase will be complete in 2003, according to CTR spokesman
Robert Poe. Project Oxygen and CTR are owned entirely by Neil Tagare, and CTR launched the
project. However, Poe says, ownership of the project has been "moved to an off-shore
corporation (Project Oxygen Ltd., Bermuda) for tax purposes."

Martin says Project Oxygen will remain a carrier’s carrier, and claims that Oxygen,
once built, will be more than a series of point-to-point hops. "We are distance- and
destination-insensitive," he says. "We’ll tell carriers, ‘We don’t care where
you come on the network or where you get off.’"

Another relatively new undersea cable player, Global Crossing Ltd., Hamilton, Bermuda,
claims to be building "the first independent, integrated global network to help
satisfy the explosive growth in demand for reliable, high-quality undersea transmission
capacity." Its Pan American Crossing (PAC) begins at Grover Beach, Calif., and goes
to St. Croix, U.S. Virgin Islands, with stops in Mexico and Panama. PAC is scheduled to
begin service in February 2000. South American Crossing (SAC) is another Global Crossing
project, aimed at linking key South American cities with PAC and the rest of the Global
Crossing network. Most of the cable will be terrestrial, but large parts of it will run
along the coasts of Argentina and Brazil. SAC is scheduled to enter service some time in
2000.

PAC and SAC, between them, amount to 26,900 kilometers of 40-gigabit, upgradable fiber
served by DWDM.

Global Crossing’s AC1 SDH undersea cable links the United States with France, Germany,
the Netherlands and the United Kingdom. The U.S.-U.K. link went into service May 1998, the
U.S.-Germany link went into service November 1998 and the U.S.-Netherlands link went into
service in January.

Global Crossing also has a "mid-Atlantic Crossing"–something of a misnomer,
since it doesn’t cross the Atlantic. The cable links New York with Bermuda, Florida,
Panama and St. Croix, and is scheduled to go into service in December.

The Global Crossing Pan-European crossing is mostly terrestrial, linking key cities in
Belgium, France, Germany, Denmark, Italy, the Netherlands, Switzerland and the United
Kingdom. But it does have two submarine links across the English Channel between the
United Kingdom, France and the Netherlands, and one across the North Sea from Germany to
Denmark. The first 13 cities will receive service in the fourth quarter of this year,
according to Global Crossing’s statements, and the final 18 will receive service in the
first quarter of 2000.

In competition with the Japan-U.S. undersea project, Global Crossing is building its
own U.S.-Japan submarine cable in the form of an SDH link between two landing points in
Japan and two in California. Partial service begins March 2000 and full service is
scheduled for June 2000.

From Clubs to Consortia

Project Oxygen, PAC and SAC, et al, are examples of privately funded cable projects.
They’re also part of huge, globe-girdling projects the authors and owners expect to
design, engineer, install, maintain and own the undersea cables when they’re in place.

MCI WorldCom and Cable & Wireless are both large, international carriers that have
chosen a traditional form–the "club"–to get across the water. Clubs, once
their cable is complete, often lease excess capacity to smaller carriers–but those
carriers and their traffic are more like ballast than passengers on the transoceanic
voyage. As regulatory barriers have come down and competitive carriers have risen around
the world, many of them have taken a new spin on the old "club" model: the
consortium.

A consortium may have dozens of members that buy into the cable, each member pledging a
certain amount of money and resources to get the cable built, each member having a say in
its building and in its operations. Depending on how much each member contributes, that
member owns capacity on the cable and can behave like an owner–leasing, or even selling,
the capacity to other carriers. Each member also has an owner’s responsibility for
maintaining the cable. The Japan-U.S. cable announced last fall has 33 members. The
members include the usual suspects–AT&T, MCI WorldCom, Nippon Telephone and Telegraph
Co. (NTT), Tokyo–but also newer and smaller players such as Level 3 Communications Inc.,
Omaha, Neb.; and Pacific Gateway Exchange Inc., San Francisco.

There is another alternative for companies that need to get across the
water–"buying capacity." What the buyer really buys is called an
"indefeasible right of use" (IRU). The owner of an IRU doesn’t own the cable and
isn’t responsible for maintaining and repairing it. However, he does have absolute control
of a certain amount of capacity, which he can use or dispose of as he wishes, for a long
time–sometimes as long as 30 years.

Lease, Buy or Build?

The argument for a carrier’s leasing, buying or building depends on its size, ambitions
and resources.

MCI WorldCom’s Blumenfeld believes consortia make sense for a project touching land in
several countries–as the Pan-American Cable project does. The cable runs from the U.S.
Virgin Islands to Aruba in the Netherlands Antilles, then to Venezuela, Colombia, Panama,
Ecuador, Peru and Chile, where it comes ashore in the northern city of Arica. Service
began on parts of the cable earlier this year, and should begin on the rest of it before
the end of the year. Although MCI WorldCom dominates the consortium with its direct
investment and its control of another investor, Sao Paulo, Brazil-based EMBRATEL, the
international long distance company, several smaller carriers also are taking part.

On the other hand, for a large carrier generating a lot of traffic, it may make sense
to build an undersea cable, Blumenfeld notes.

"If you have significant enough … requirements for capacity, it might be worth
it to build your own (undersea cable), which you control," Blumenfeld says.
"Then you don’t need the votes of members of a consortium to do various things. For
MCI WorldCom, since we were developing operating companies in western Europe, we reached
the conclusion that we could afford to build our own, which we keep essentially for our
own use, though we do lease some capacity."

Not surprisingly, Global Crossing and Project Oxygen think the only thing to do is
lease capacity on either of their huge projects. It’s important here to understand the
difference between leasing and buying capacity. A carrier leasing capacity on an undersea
cable is entitled to a finite amount of capacity for a set price and a specific–usually
relatively short–time period. Such a carrier doesn’t own the cable or the optical fiber
in it, and therefore may have to accept some restrictions on how it’s used. To "buy
capacity" really means to buy an IRU. Under an IRU, a carrier doesn’t own the cable,
but does have absolute control over a certain amount of capacity in it for a relatively
long period. Such a carrier’s lot is somewhat analogous to that of a condominium owner who
owns the interior walls of his house, but not the exterior. And just as the condo owner
doesn’t have to worry about repairing the roof or cutting the lawn, neither does the
holder of an IRU have to worry about maintaining or repairing the cable.

Viatel Inc., New York, which is building a pan-European fiber optic network in Europe
called Circe, uses IRUs and leases to get across the Atlantic. Circe has some undersea
links in the English Channel and the North Sea, but the deep-water glass is mostly
acquired through IRUs.

"We came to the decision that our margins would be better if we got rid of static
leased costs," says Glenn Davidson, Viatel’s vice president-corporate communications.
"That way, we’d have better control of our costs and of quality. What we’ve done is
purchase or sometimes lease capacity on trans-Atlantic cable."

Circe is being built partly with dark fiber, Davidson says, so Viatel can swap it for
needed trans-Atlantic capacity in the future.

Ken Branson is business and finance editor for PHONE+.