While transatlantic and pan-European cable construction seems to have ebbed after reaching a saturation point in the last year or two, the action now seems to have shifted to the trans-Pacific and pan-Asian domain.
Since around the end of last year, there have been a number of subsea cable landings and service launches in Asia, such as the completion of the redundant leg of Asia Global Crossing's Pacific Crossing-1 trans-Pacific cable.
However, the same time period also saw select subsea cable consortia in high-profile repair mode. Earlier this year, the China-US Cable -- the first trans-Pacific cable directly connecting the US and China -- suffered a major blow as a section off the coast of Shanghai was snapped in two. The event cut off the majority of China's overseas 'Net connectivity and forced carriers with capacity booked on the system to reroute the traffic on to other systems.
Less than a month later, fishing trawlers snapped another section of the China-US cable -- this one a coastal underwater link connecting Shanghai and Shantou, the system's two landing points in China.
The second cut had less of an overall impact on regional traffic flows compared to the first one, but both episodes serve to illustrate that the business of building subsea cables is easy to take for granted -- especially when considering that it's been 153 years since someone first deployed a communications cable underwater.
Obviously, the business has come a long way since then in terms of technology and cost, but after a century and a half of trial and error, the actual act of deploying a subsea cable network has become a fairly straightforward procedure.
As Asia Global Crossing's MD for network development, David Milroy, puts it, "Deploying a submarine cable system isn't rocket science."
Of course, he adds, that doesn't mean it's child's play, either. Indeed, there is far more to deploying a fiber-optic cable system under a few kilometers of water than spooling it out off the back of a boat. In fact, comparatively speaking, that's the easy part.
Choose your armor
According to AGC's Milroy, just the initial planning of a subsea system involves a lot of legwork even before the ships start rolling out fiber into the sea.
"Before you even think about putting a cable system underwater, you have to take a number of things into account, such as the network engineering, what it's made of, where it goes, how it's manufactured," Milroy says. "For example, do you need single or double armor, and for which sections do you need it?
"Typically you'll need a double wrap [armor] for shallow water up to about 1,000 meters," explains Milroy, citing fishing trawlers and ship anchors as the most common hazards to be found at those depths. "If the water is deeper than that, you're pretty safe with single armor, since fishing nets and anchors typically don't reach that far."
K.F. Larm, regional director for Level 3's global submarine division, says that lightweight cables are okay for depths below 2,000 meters, but recommends light armor at depths of around 1,500 meters, where the danger isn't so much anchors and trawler, but sharks.
"If the cable becomes free-floating for some reason and isn't lying flat, the current will move it back and forth, and sharks will try to bite the cable."
(Those who wonder just how much damage a shark could possibly do to a fiber-optic cable are hereby reminded that (a) the optical fiber inside the cable is essentially made of glass, and (b) sharks are known to have a biting torque that measures in the range of several hundred pounds per square inch.)
Another reason for determining armor requirements in advance is that the type of armor used affects how much cable can be carried by a ship at one time.
Captain Frank Kitt of the CS Bold Endeavour -- part of the Global Marine fleet owned by Global Crossing - says that the two storage drums in the hold of his ship can carry between 4,000 and 6,000 km of fiber, depending on the armor thickness. "Obviously, a double-armor cable is thicker than single-armor -- about 46 mm compared to 31 mm," says Kitt. "That may not sound like a big difference, but it is when you coil it up in one drum."
Survey says
Another aspect of pre-planning a submarine system, says AGC's Milroy, is mapping out the undersea terrain.
"You have to do a marine survey, where you're taking a look at the sea bed and looking out for things like mountains and valleys and areas where earthquakes might pose a problem," Milroy explains. "You want to avoid natural and man-made formations and areas where there's plate movement activity. At the same time, you want the cable path to be as straight as possible. You don't want any bends because you risk breaking the fiber."
Even more potentially challenging are the things that most people outside the international carrier business might not associate with subsea systems at all, says Level 3's Larm, who offers a fairly long checklist of items.
"Oil exploration and exploitation areas, military zones, dumping zones, third-party territorial waters and political claim areas are all things to watch out for," Larm says. "You also want to avoid areas where there's heavy fishing activities or dredging. You also want to take things like water currents into account."
Then there's the matter of selecting a landing site. This is not to be confused with getting a license to land an international cable in a market -- which is certainly a key issue, but only for private systems like East Asia Crossing and Level 3's Tiger network, since club cables like SEA-ME-WE-3, APCN-2 and FLAG are the products of incumbent telcos who have had permission to land cables domestically since at least the start of the 20th century.
According to Su Weichou, VP for the Greater China market for Level 3, there are many issues to consider in deciding where to physically land a cable. He suggests keeping the cable away from the local shipping channels. "The reason is that you have to consider your future operation and maintenance plan," he says. "If you run the cable through a shipping channel and it breaks, you have to stop all ships from using that channel while you're repairing the cable."
Su offers Level 3's own experience with selecting landing sites for the Tiger system as an example. "In Taiwan, we couldn't land cable anywhere on most of the east coast because there's a lot of undersea volcano and earthquake activity just offshore," Su says. "Landing the cable on the southern tip presented a backhaul problem because it's too far away from the exchange. We also had to be careful with deploying cable in the Taiwan Straits, because it's a politically sensitive area."
By the book
The bright side is that by the time all of that gets sorted, deploying the cable itself is actually a relatively by-the-book affair. Once the cable itself has been manufactured according to spec, it's loaded onto the ship.
"After the shore ends are put in," says AGC's Milroy, "at both ends you run about 10 to 15 km of cable from the [beach] manhole and tie it off to a buoy, then you basically run the cable between the ends according to the survey."
That survey is loaded into the ship's computers, so the ship is in essence preprogrammed to follow the undersea route plotted out for the cable.
During this process, the cable isn't just lying on the sea bed -- at least not for the whole length of the system, although this was standard practice until at least the 1960s. Nowadays, the cable is buried under the sea bed, offering further protection from fishing boats, anchors and shark attacks.
Burial depth varies according to the situation. Burying the cable becomes optional at depths below 2,000 meters. For those who choose to bury the cable at those depths, a meter under the floor is usually sufficient. In shallower waters (from 1,000 meters and up), burial is virtually mandatory, and a burial depth of 10 meters is strongly recommended, especially as the system gets closer to shore.
The burial process is performed by the self-explanatory cable plow, which is remotely controlled from a bay onboard the ship. So is the ROV (remote operated vehicle), a track-mounted vehicle whose job is to guide the cable to the appropriate place mapped out for it on the sea bed. Both are bristling with video cameras so the remote operators can see the floor and make sure the burial is done properly. The cable plow is typically capable of plowing through solid rock as well as sediment.
Naturally, this is painstaking work -- the ship's speed will usually average between 1 to 2 km per hour, slowing down every 50 km or so when a repeater is about to go over the side -- one reason why cable laying is a 24-hour-a-day operation.
Bandwidth on demand
Despite the massive amount of preparation involved in planning out a subsea cable and the fairly routine nature of deploying it, things do go astray in the execution. Indeed, how to explain the fact that modern technology has given us double-layer armor and the ability to bury cables 15 meters deep under two kilometers of water, yet 20 million Internet users in China can be cut off from every single overseas Web site on the 'Net courtesy of a fishing boat?
Some industry analysts have observed that both cuts occurred near China in areas that are heavy traffic areas for fishing boats that use net anchors heavy enough to sink well below a meter into the sea floor, especially when they are dragged along the bottom. As such, no one could expect a fiber-optic cable to last long in those areas unless it was buried good and deep.
The real shocker in the case of the China-US cable, of course, was that it shouldn't have mattered where the cable was cut, as the system was designed as a redundant loop that would provide instant failover to backup capacity.
That didn't happen because the second leg of the loop hasn't been finished yet because of environmental issues raised in the system's second landing point in San Luis Obispo, Calif., which means that the China-US cable has been operating with no redundancy protection since its service launch in January 2000.
This isn't an unusual thing. The business logic is simple: Why wait for the rest of the system to be operational in 12 months when you could be spending that time selling what capacity you have and earning revenue off the traffic volumes? Essentially, fiber operators are gambling that a fiber cut won't happen until long after the rest of the system s built. The China-US cable club lost that bet, although to be fair, if construction had stayed on schedule, Internet users whose packets were being routed on that system might never have noticed any difference in service quality.
Interestingly, the concept of Sonet/SDH protective rings has been around for over 10 years, yet most subsea cable systems built over the last decade are not built on redundant topologies.
"There are a surprising number of single-shot cables, and other carriers are just now realizing the vulnerability of those systems," Milroy says.
He doesn't mince words when offering his thoughts on why carriers would willingly skimp on redundancy. "The reason no one does it is because it costs more money," Milroy says candidly, who adds that redundancy is a figment of the imagination of many carriers.
"What most carriers do is they buy capacity on multiple cable systems as backup -- the problem is that if a cable actually fails for whatever reason, they have to go through the process of rerouting the traffic from that cable to the other cables, which is why it takes them 10 hours to do it."
RELATED ARTICLE: 153 years of submarine cable history: the highlights
1848 First submarine cable laid across the Hudson River in New York
1858 First transoceanic subsea cable deployed (and destroyed the same year)
1866 The world's first working transatlantic telegraph cable is launched, with a transmission speed of eight words per minute
1867 Invention of siphon recorder speeds up subsea telegraph transmission to 10 words per minute
1870s Duplexing becomes widely used.
1880 Transatlantic subsea traffic now averages 1,500 messages per day in each direction
1902 The first trans-Pacific cable is completed between the US and New Zealand
1955 The world's first amplified cable, TAT-1, is deployed across the Atlantic
1988 TPC-3, the first fiber-optic trans-Pacific cable, goes online
Source: Pacific Telecommunications Review, 3Q00
RELATED ARTICLE: FIX THIS! The art of cable O&M
Pop quiz: You are the owner of a multi-billion-dollar submarine fiber-optic cable system. A fishing trawler's net snags one of your cables and drags it until it snaps, causing a traffic outage that is costing you millions of dollars a minute. What do you do?
If your system is built with a redundant architecture, the one thing you don't do is panic, says Asia Global Crossing's David Milroy. "If you've got a ring system like we have, you're not so pressured to get it fixed."
Pressure or none, the first step is to determine the physical location of the break. "We use differential GPS to keep track of where the repeaters are, as well as the cable itself," says Milroy. "So if there's a break in the cable, we can find it. Our network operating center can give us the exact location by tracking how far the impulses go before they stop, and the GPS gives the maintenance ship the exact physical location on the surface."
The next step is to notify the O&M provider, who keeps a dedicated fleet of ships ready to go for just such an occasion. Of course, it's impossible to know where a cable cut might happen, but Christian Reinaudo, president of the optics division of Alcatel -- which provides O&M services with its fleet of 12 ships -- says it is possible to make an educated guess.
"Usually cables get cut close to shore, fortunately, so you do have a fair idea of where it's most likely to happen," Reinaudo says. "So we locate our ships close to those risky zones."
Once the ship arrives on the scene, which can take as long as a few days, depending on the location of the break, the crew drops the subsea cable ship equivalent of a tailhook into the water, which catches the cable. The severed ends of the cable are then pulled to the surface and hung on buoys, AGC's Milroy explains, with the cut ends brought onboard and spliced to a new length of cable that will run between them. "Obviously it's impossible to splice the original two ends back together, because the cable is pretty taut to begin with," he says.
COPYRIGHT 2001 Questex Media Group, Inc.
COPYRIGHT 2008 Gale, Cengage Learning
