Copyright 1998 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.
This article was published in the February 1998 issue of
IEEE Communications Magazine.

Technical Paper

Abstract

WDM technology is now being applied to international undersea fiber optic cable networks in order to provide enhancements such as increased network capacity and greater network flexibility. This article looks at what WDM technology can provide, the progress being made, and the special challenges in its application in undersea networks. We then describe several international undersea networks that, when completed by the end of 1999, will use WDM technology and will serve as a major part of the global undersea fiber optic infrastructure connecting the world.


Applying WDM Technology to Undersea Cable Networks

Patrick R. Trischitta and William C. Marra
Tyco Submarine Systems Ltd. Laboratories

The same market forces that are driving the need for the enormous expansion of local, regional, and national telecommunications networks are also driving the need for extraordinary expansion of international networks. Global forces such as the expanding Internet, the continuing worldwide liberalization of telecommunications markets, and the continuing globalization of business and trade have created unprecedented demand for larger, longer, more complex international undersea cable networks [1].
To meet this demand, undersea network providers are now using wavelength-division multiplexing (WDM) technology to design networks with capacities and capabilities that were considered unimaginable just a few years ago. WDM technology is now having as dramatic an effect on the planning and deployment of undersea cable networks as it is on the rapid expansion of terrestrial fiber optic networks.
This article looks at the progress being made in applying WDM technology to undersea cable networks. We examine what WDM technology provides, how it can be applied to the undersea environment, and what special challenges exist in its application. Then we describe several international undersea networks that will use WDM technology to enhance network capabilities. These networks, when completed by the end of 1999, will serve as a significant new piece of the fiber optic global infrastructure connecting over 90 countries with 100,000 km of undersea cable.

What WDM Technology Provides to Undersea Networks

Not unlike their terrestrial network counterparts, WDM technology provides undersea fiber optic networks with two fundamental enhancements: increased capacity and greater network flexibility. Both enhancements are now being used to continue the evolution of undersea networks from point-to-point fiber optic links to highly sophisticated networks. These networks provide branches that allow connectivity to extend to more countries, and automatically restore traffic in the event of an accidental cable cut. This evolution is being driven even faster than expected by a seemingly insatiable global demand for direct large-bandwidth connectivity between more countries, as well as the need for this connectivity to be available virtually 100 percent of the time.

Increased Capacity

When used with undersea repeaters containing erbium-doped fiber amplifiers (EDFAs), WDM allows greater capacity per fiber than could be realized using a single wavelength carrier. This is having a dramatic effect not only on the design capacity of new cable systems but also on optical-amplifier-based undersea systems installed since 1994. By adding shore-based terminal equipment tuned to different wavelengths, the transport capacity of these existing undersea cable networks is being doubled and in some cases tripled. For example, TAT-12/13 [2] and TPC-5 [3] cable networks, in-service since 1996 and originally designed and installed to operate with a single carrier channel at 5 Gb/s, are now being upgraded with a second and then third wavelength in the case of TAT-12/13, with each wavelength providing an additional 5 Gb/s of capacity [4]. These in-service capacity upgrades are extending the usefulness of these networks at a time when capacity demand exceeds all previous estimates.
However, it is in the design of new undersea cable networks where the full advantages of WDM technology can be realized. Transoceanic systems with up to sixteen wavelengths per fiber are now possible over distances of 8,000 km. These systems transport an STM-16 (2.5 Gb/s) on each wavelength, supporting a maximum capacity of 40 Gb/s per fiber. Also possible is eight channels at 5 Gb/s over similar distances. The key challenge in the design of these systems is how to achieve a large number of wavelengths over distances as large as 12,000 km. Special care must be taken in choosing the dispersion map of each fiber path and in the spacing of the wavelengths [5]. In the future, this may increase to an STM-64 (10 Gb/s) per wavelength.
In the laboratory, system experiments have been done showing the feasibility of up to 32 channels at 5 Gb/s per carrier over 9300 km [6] and up to 16 channels at 10 Gb/s over 6000 km [7]. On a four-fiber-pair cable this would be a capacity of 640 Gb/s. However, significant further development is required to take these laboratory results to reliable undersea networks.

Network Flexibility

The second fundamental enhancement that WDM technology provides is in networking capabilities. By adding an optical wavelength layer to the system where traffic routing can take place, network functionality is no longer constrained by the number of physical fiber paths in the undersea cables. Wavelengths within the fiber paths can be routed and reconfigured independently.
This enhancement creates an opportunity to design more complex undersea networks with more landing points and more flexibility in traffic routing with fewer fiber pairs and the corresponding undersea equipment. This is particularly advantageous to undersea networks where the number of fiber pairs in the network is limited by the physical constraints of the undersea cable and the design of the repeater housing. Contrast this with terrestrial networks, where it is the cost and availability of the right of way that is a major factor. In terrestrial networks once a right of way is obtained, it makes a lot of sense to install as many fiber pairs in the cable as practical, say 48, 96, or even more. Repeater stations and terminal stations can then be equipped with only the repeaters and terminals needed for initial service, with additional equipment installed as needed.
For transoceanic cable systems, international waters provide free rights of way. However, system costs are related to a number of key factors, including the number of amplifier pairs in the undersea repeaters. Today, all suppliers of transoceanic undersea networks manufacture repeaters with up to four pairs of optical amplifiers. Repeaters that support greater than four amplifier pairs are possible, but for economic and technology reasons adequate capacity has been provided by systems which support up to four fiber pairs.
WDM allows undersea networks to use the wavelength layer to add and drop more traffic capacity at more landing points, while keeping the number of fiber pairs in the system to a minimum. This feature is a result of adding wavelength-selective filtering capabilities to undersea branching units (BUs).
Some of the additional benefits provided by the optical wavelength layer include:
  • Undersea or shore-based wavelength add/drop capability
  • Flexible traffic routing
  • Reconfigurable add/drop, allowing for undersea bandwidth management and restoration capabilities.

Special Requirements for Undersea Networks

A principal requirement for all undersea networks is very high reliability, in particular for the portion of the network that is undersea. Redundant components are used where necessary to ensure network availability. While undersea networks can rely on alternate facilities (typically other cable facilities or possibly satellites) for traffic restoration in the event of catastrophic failures due to external aggression (e.g., a ship anchor cutting a cable), undersea cable traffic capacity is getting too large to be economically and reliably restored via satellite. This has driven the need for in-network cable-on-cable restoration, and with this, the first deployment of large self-healing ring networks (e.g., TAT-12/13 and TPC-5).
The trend for in-network cable-on-cable restoration continues for networks using WDM technology. Figure 1 shows the two major cable topologies found in undersea fiber optic networks that provide in-network restoration features. Ring networks provide nearly instant restoration of traffic in the event of a cable cut. Another popular undersea network topology is trunk and branch. This type of network topology typically provides direct connectivity to many countries and includes a main trunk cable in deep water, off the continental shelf, and branch cable connections to the shore via undersea BUs [8]. These networks form self-healing rings (bidirectional line-switched rings), within a linear cable. At first glance, restoration on the cable itself may seem unusual. However, most external aggression events that sever cables occur in shallow water. Therefore, these self-healing networks maintain traffic between all countries on the network except that country eliminated by the cable break in the branch cable.
In the next section we will describe how these enhancements afforded by using WDM technology are being used in systems now under construction.

Planned Undersea Networks Using WDM

Africa ONE

Originally conceived to solve the telecommunications problems in Africa in one giant leap, the Africa ONE project was the first undersea network planned to use WDM technology [9]. Although the system has never come to contract, Africa ONE nevertheless provided the technological framework that showed the potential of WDM technology in undersea networks (Fig. 2). Originally conceived to encircle the continent of Africa with a 40,000 km ring of undersea fiber optic cable, this system was to use eight wavelength channels on each of two fiber pairs. Between central offices were to be up to eight undersea wavelength add/drop multiplexes, where a single wavelength at 2.5 Gb/s was dropped from and added to branch cables. In this way every coastal country of Africa was to be connected to the network.

SEA-ME-WE-3

Extending on the design concepts developed for Africa ONE, the SEA-ME-WE-3 Cable System will be the first undersea network to use WDM technology to do undersea routing of wavelengths. Scheduled to go in service in late 1998, SEA-ME-WE-3 will use undersea wavelength add/drop multiplexing to realize a complex traffic connectivity over two pairs of undersea fiber. Figure 3 shows the route of the SEA-ME-WE-3 cable, which extends from Germany to Singapore.
This network will be installed with a capacity of up to eight wavelengths on each fiber with each wavelength carrying an STM-16. SEA-ME-WE-3 will be the first undersea network to use WDM in a multipoint undersea wavelength add/drop configuration. Connecting more than a dozen countries between Southeast Asia (SEA), the Middle East (ME), and Western Europe (WE), this network has a trunk and branch cable topology and uses undersea wavelength add/drop multiplexing BUs. Wavelength-selective filters in the undersea BUs add and drop specific wavelengths to and from the individual branch cables. This allows efficient allocation of the full capacity on each individual fiber pair to separate countries, either directly on the network or indirectly connected via transit facilities. This configuration results in a very high degree of traffic sovereignty and security.
To illustrate how the wavelength layer will be used for traffic routing, Fig. 4 shows the wavelength connectivity of four wavelengths on each of the two fiber pairs on the cable between Djibouti and India.

Atlantis-2, Columbus-3, and Americas-II

These three separate WDM systems, shown in Fig. 5, when completed by the middle of 1999 will form a ring around the South Atlantic Ocean connecting countries in four continents. Unlike Africa ONE and SEA-ME-WE-3, each of these systems will use shore-based wavelength multiplexing and demultiplexing. This allows for the undersea BUs to be optically passive, and the capacity and traffic routing to be changed via the shore-based terminal equipment after the system is operational. By stacking the wavelengths on each fiber path (up to eight wavelengths on each fiber pair), capacity of the system can be increased as traffic demand grows. This reduces the initial system cost since fewer wavelengths can be equipped initially.

China-US and Atlantic Crossing -1

Even though the first trans-Atlantic and trans-Pacific ring networks, TAT-12/13 and TPC-5, were just completed in 1996 and are scheduled to be upgraded to two or three wavelengths on each fiber in the next few years, demand for capacity across the Atlantic and Pacific outpaces supply. To fill this demand two new transoceanic ring networks will be deployed. The China-US Cable System and the Atlantic Crossing-1 Network (AC-1) will be built using the most advanced WDM technology available today to achieve the maximum capacity possible over the longest distances. Both systems are WDM stacked ring networks that carry traffic capacity on four fiber pairs. China-US is designed to carry eight wavelengths at 2.5 Gb/s over the maximum distance of 12,000 km (Fig. 6), whereas AC-1 is designed to carry its traffic capacity over a maximum distance of 7100 km. The AC-1 ring, shown in Fig. 7, connects four landing points using four separate undersea cable segments: U.S.-U.K., U.K.-Netherlands, Netherlands-Germany, and Germany-U.S. Each segment will contain four fiber pairs, and each fiber will initially carry four wavelengths at 2.5 Gb/s. Thus, the initially deployed transmission capacity will be 40 Gb/s per segment. Although there are four landing points, the network is constructed to form two self-healing ring networks, with each ring network consisting of eight WDM channels. The logical implementation separates the STM-16 WDM channels into service and protection on separate fiber pairs. The resulting architecture of each physical ring is four independently operating four-fiber ring networks, each consisting of three nodes. The first ring includes network nodes at the United States, United Kingdom, and Germany, and the second ring includes network nodes at the United States, Netherlands, and Germany.

Capacity Expansion Trends

The rapid progress and enormous capacity-expanding capabilities of WDM technology can be highlighted by looking at the historical capacities provided by trans-Atlantic fiber optic cable systems. TAT-8 was the very first trans-Atlantic cable system to use fiber optic technology. Going into service in 1988, TAT-8 carried 560 Mb/s of capacity on two fiber pairs. In 1996 the TAT-12/13 ring network was completed with 32 times more capacity. By using optical amplifier technology in the undersea repeaters, TAT-12/13 transports a single wavelength channel at 5 Gb/s on each of four fiber pairs separated into two diversely routed cables. TAT-12/13 is already under contract to be upgraded using WDM technology, first to 10 Gb/s per fiber pair by adding a second wavelength channel at 5 Gb/s, and then to 15 Gb/s per fiber pair by adding a third wavelength channel on the existing four fiber pairs. New trans-Atlantic systems, such as AC-1, are expected to reach capacities of at least 20 Gb/s (8 x 2.5 Gb/s) on each of eight fiber pairs spread over two diversely routed cables. The resulting capacity carried across the Atlantic will be astonishingly 250 times larger on these WDM networks than that carried on the TAT-8 system. This growth rate represents a doubling of circuit capacity across the Atlantic Ocean every 18 months.
Further increases in system capacity are expected with the use of more advanced WDM technology. Laboratory experiments have demonstrated 160 Gb/s capacity per fiber, allowing 640 Gb/s of capacity in a system with four fiber pairs, over a system length of 9300 km using 32 wavelengths each carrying 5 Gb/s. The same capacity of 160 Gb/s per fiber pair over 6000 km has also been demonstrated using 16 wavelengths at 10 Gb/s. Significant further development is required to take these laboratory results to reliable undersea products, but in the first decade of the next century, networks with capacities approaching 1 Tb/s will be constructed to meet the international traffic demands of the 21st century.

Undersea Networking Trends

Similar extensions in capacity could be applied to trunk and branch networks. Additional wavelengths associated with higher-capacity systems would increase the benefits of wavelength add/drop multiplexing. Future undersea branching units might need to select one or more channels to be dropped from any of 10–40 wavelengths, depending on total capacity and the bit rate of the individual channels. Thus, higher levels of optical integration or other innovative technologies may be required.
In addition to extending the current network architectures to higher capacities, it may make sense in the future to consider other, more innovative undersea architectures. Branching units using WDM switching to perform network protection or traffic reconfiguration are attractive alternatives. Undersea networks composed of many separate wavelengths with all-optical routing, including failure protection and traffic reconfiguration, offer the opportunity for more reconfigurable capacity to more landing points than ever before.

Summary

WDM technology is already having a profound effect on the amount of capacity that can be transported across the world's oceans. The systems described, when completed, will become the major backbone networks for global communications.
The advancement of WDM technology will undoubtedly continue to have a major impact on the design of international undersea cable networks. As the technology develops to allow for more wavelengths to be carried longer distances and at higher bit rates, new undersea cable networks will be built connecting more countries with more capacity than ever before.

Acknowledgments

The authors wish to acknowledge the many engineers, scientists, and managers at Tyco Submarine Systems Ltd. whose innovative talents have developed many of the concepts and technologies used in the networks described in this article.

References
[1] P. R. Trischitta and W. C. Marra, Special Topic on Global Undersea Communication Networks, IEEE Commun. Mag., Feb. 1996.
[2] P. R. Trischitta et al., "The TAT-12/13 Cable Network," IEEE Commun. Mag., Feb. 1996.
[3] W. C. Barnett et al., "The TPC-5 Cable Network, " IEEE Commun. Mag., Feb. 1996.
[4] J. Feggeler et al., "WDM Transmission Measurements on Installed Optical Amplifier Undersea Cable Systems," Proc. OFC '96.
[5] N. S. Bergano and C. R. Davidson, "Wavelength Division Multiplexing in Long-Haul Transmission Systems," J. Lightwave Tech., vol. 14, no. 6, June 1996.
[6] N. S. Bergano, et al., "Long-Haul WDM Transmission Using Optimum Channel Modulation: A 160 Gb/s (32 x 5 Gb/s) 9300 km Demonstration," Proc. OFC '97, PD16.
[7] N. S. Bergano, et al., "Long-Haul WDM Transmission Using 10 Gb/s Channels: A 160 Gb/s (16x10 Gb/s) 6000 km Demonstration," Optical Amplifiers and Their Applications 1997, PD9.
[8] P. R. Trischitta, R. Cofre, and A. Medina, "The Pan American Cable System," IEEE Commun. Mag., Dec. 1997.
[9] W. C. Marra and J. Schesser, "The Africa Optical Network," IEEE Commun. Mag., Feb. 1996.

Biographies
Patrick R. Trischitta [F] is a technology director in the System Design Group at Tyco Submarine Systems Laboratories in Holmdel, New Jersey. He received B.S.E.E. and M.S.E.E. degrees from Georgia Tech and a Ph.D. in electrical engineering from Rutgers University. He joined the Undersea Systems Development Laboratory of AT&T Bell Laboratories in 1980. Since that time, as a member of technical staff and then distinguished member of technical staff, he has worked on the design and development of fiber optic telecommunication systems mostly for undersea applications. He has served as an associate editor of IEEE Transactions on Communications and the IEEE/OSA Journal of Lightwave Technology. He presently serves as chair of the Journal of Lightwave Technology Steering and Coordinating Committees. He is presently responsible for the system design of several transoceanic undersea communication networks.
William C. Marra is director of technical marketing at Tyco Submarine Systems Laboratories. He received his Bachelor's degree in electrical engineering from Polytechnic Institute of Brooklyn, a Master's degree in electrical engineering from Stevens Institute of Technology, and a Ph.D. in electrophysics for joint work done at Stanford University and Stevens Institute. He joined the basic research organization of AT&T Bell Laboratories in 1969. As a member of technical staff and then distinguished member of technical staff, he has worked on the development of numerous fiber optic telecommunications systems for both terrestrial and undersea applications. Before assuming his current position, he was the chief architect responsible for deploying transoceanic ring technology to the TPC-5 and TAT-12/13 cable networks, and WDM technology for Africa ONE, and has guided the network design efforts on the first commercial application of WDM technology to the transatlantic ring network, Atlantic Crossing-1.