Laboratory work stretches transoceanic reach and capacity limits
By ROBERT PEASE
Using dense wavelength-division multiplexing (DWDM), Alcatel (Paris) successfully transmitted 10 Gbits/sec over 32 channels on a single fiber across a distance of 6150 km, achieving an overall capacity of 320 Gbits/sec per fiber. The laboratory experiment, conducted on a 4-fiber-pair system, represents a four-fold increase in today`s undersea network capacity.
Claiming the first reported demonstration of 32䁾-Gbit/sec transmission over transoceanic distances, Alcatel`s design teams are preparing for the launch of new products applicable to future high-capacity submarine networks.
"This is the first demonstration of a practical way to achieve 320 Gbits/sec for transoceanic distances," says Paul Gabla, director of central marketing, Alcatel Submarine Networks. "There is a race for more capacity in international and transoceanic systems, both terrestrial and submarine."
Gabla says the capacity of operational systems, as well as those being installed today, will be as high as eight channels at 2.5 Gbits/sec, a total of 20 Gbits/sec, over two fiber pairs. That equates to an aggregate capacity of 40 Gbits/sec, which is considered state-of-the-art for both existing and in-progress systems. But with the explosion of the Internet and the demand it places on networks, carriers around the world are struggling to find ways of achieving more capacity from their fiber.
Even though submarine systems under construction today are not being designed to operate at these huge capacities, they are being built with scalability in mind. For example, a Japan/U.S. system being constructed by an international 33-company consortium (see Lightwave, September 1998, "Consortium inks Japan/U.S. undersea cable construction deal," page 23) will initially provide a capacity of 20 Gbits/sec per fiber pair. With future growth in mind, current construction is expected to allow its eventual growth to 160 Gbits/sec on each fiber pair.
"In a submarine fiber-optic system, you need to know from day one where you want your system to go," says Gabla. "You have to be sure the submerged cable and repeaters will be able to support the ultimate capacity that you want."
With a better knowledge today of optically amplified, multichannel systems, Gabla says new networks can be designed with a more accurate margin assessment for future upgrading. For instance, a currently operating network designed for 16 wavelengths at 10 Gbits/sec may be able to upgrade to 32 wavelengths as technology matures.
Gabla has observed that carriers appear to be rushing to buy capacity even though they aren`t using it yet. Such actions may not be so strange, though, when you consider most conservative forecasts say that traffic demand, spurred by Internet growth, will continue to more than double each year.
"Carriers may not use the capacity now, but they`re anticipating using it very soon," says Gabla. "Given the time it takes to implement a new system, if they don`t secure it very soon, they could be in big trouble."
New considerations
Increasing capacity in systems that span miles of ocean presents unique problems for transoceanic systems engineers. The equipment used in submarine systems must be ultra-reliable, since the system is designed to last 25 years. Unlike terrestrial systems, where a failure can be handled by simply replacing failed equipment, maintenance on an undersea system is extremely difficult and expensive. A more obvious issue is the need to transmit over long distances, as much as 9000 to 10,000 km, without regeneration. A terrestrial system rarely exceeds 4000 km, which is about the distance of a transcontinental system across the United States.
With a formula of growing demand, expensive capacity, maturing technology, and possibly limited availability of deployed fiber, it`s little wonder transoceanic carriers are feeling pressure to squeeze the most capacity possible from systems designed for operation through several decades. For undersea systems in place, upgrades may be limited with existing equipment. To achieve higher bit rates and more wavelengths, elements of the system will need modification.
According to Gabla, there are three areas of focus in research efforts to upgrade undersea systems. The first area to consider is the fiber itself. The new large-effective-area fiber currently being marketed by companies like Corning Inc. (Corning, NY) and Lucent Technologies (Murray Hill, NJ) are examples of work in this area. The fiber is not the main focus in most labs, says Gabla, but improvements in fiber performance could help reduce repeater spacing, decreasing the number of repeaters required in an undersea system.
Improvements can also be made in the area of repeaters, says Gabla. Flattening the bandwidth can allow more channels to be stacked in the amplified line, and more channels equate to more capacity. Experiments have also been conducted toward using channel spacing narrower than has been used so far to stack more channels in the same bandwidth. For example, Alcatel has used a 0.4-nm channel spacing to put 32 channels in a system designed for 16.
Using more powerful pump lasers to increase output power is another area for improving repeater performance. However, says Gabla, the technology is fairly immature and not much improvement has been gained so far in this area. The area of concentration for Alcatel is improved signal processing in the terminal equipment, on both the transmit and receive sides.
"Signal processing means better controlling the shape of the transmit pulses," says Gabla. "By applying some clever algorithms, you can more efficiently demodulate the signal at the other end. We`re working on this at Alcatel, but a combination of success in all these areas will eventually lead to solutions for real-world long-distance systems as opposed to laboratory experiments."
Others have reported success in expanding capacities in submarine and terrestrial systems. Tyco Submarine Systems Ltd. (Morristown, NJ), for example, transmitted 64 channels at 5 Gbits/sec, achieving the total capacity of 320 Gbits/ sec. But Gabla points out that 5 Gbits/sec is not a standard bit rate and carriers don`t support it because of some very basic interface problems when connecting to terrestrial systems.
"This is not the first time transoceanic distances have been bridged," says Gabla. "It`s not the first demonstration of 320 Gbits/sec. But it is the first time the two have been demonstrated over these kinds of distances with 10-Gbit/sec channels. That`s why I believe it`s closer to a real application." q