Broadband network increases transport capacity and reduces cost via dense WDM

A broadband transport network infrastructure combines an eight-channel dense WDM scheme, optical amplifiers, nonzero-dispersion fiber and end terminals to cost-effectively boost transmission capacity to 20 Gbits/sec

OVE Parmlind

Lucent technologies

The broadband transport network (BTN) uses innovations developed at Bell Laboratories to eliminate the need for numerous optical fibers and regenerators or single-channel optical fiber amplifiers in new high-capacity networks. Fewer components result in lower network costs and improved network efficiency and reliability. The solution also allows service providers to increase at least eight-fold the capacity of their existing fiber-optic networks. The BTN consists of three main elements: Truewave singlemode optical fiber; optical line system (OLS) terminal equipment and optical amplifiers; and Synchronous Optical Network (Sonet) and Synchronous Digital Hierarchy (SDH) transport terminals.

Additional benefits of the broadband network include future-proofing networks by allowing capacity to be added in increments as demand grows; increased network reliability by combining dense wavelength multiplexing and self-healing ring capabilities; and a dedicated ninth wavelength for integrated maintenance message communications. The dense wavelength multiplexing technology works cost-effectively in both 2- and 4-fiber-line bidirectional switched-ring architectures.

Regenerator savings

Dense wavelength multiplexing decreases outside plant cost by reducing the number of electro-optical regenerators used. For example, a typical 360-kilometer, 20-Gbit/sec, dense wavelength-division-multiplexed-based network consists of 64 regenerators to support a conventional 8-fiber parallel transport network (see Fig. 1). In this setup, eight 1310-nanometer regenerators are needed at every 40-km spacing to amplify the optical signals.

With the BTN approach, a single fiber supplants the operation of eight fibers. Two OLS end terminals are used to process the eight wavelengths. Between the end terminals, only two optical repeaters are needed--at the 120- and 240-km distance markers--to amplify the signals.

In contrast, the conventional 1300-nm network requires the use of eight electro-optical regeneration stations, one at every 40-km marker, for each wavelength. The BTN approach therefore eliminates 62 regenerators and seven fibers, which represents a sizable capital savings in space, equipment, electronics, utility, operation and maintenance expenses.

In the fiber-optic backbone networks installed during the past 15 years, many network planners underestimated their future communications capacity needs. They often specified the installation of cables containing a limited number of fibers. Presently, these networks experience a shortfall in capacity because of the explosion of customer demands for expanded voice, video and data traffic. In fact, in numerous networks the demand for increased data traffic has exceeded that for voice.

To overcome capacity problems on long-distance fiber networks, service providers have implemented various solutions. One approach was to increase the bit rate at the end terminals via updated electronics. These bit rates have increased steadily every few years--from 45 megabits per second in the late 1970s, to 156 Mbits/sec in the 1980s, and to 2.5 and 10 Gbits/sec in the 1990s. To generate these higher rates, service providers relied mainly on installing faster-operating laser devices.

This quest for faster bit rates also resulted in another approach for increasing network capacity--wavelength-division multiplexing (WDM). This technology processes multiple wavelengths over a single fiber. This approach contrasts with adding more and expensive single-fiber transmission channels. Dense WDM generally refers to implementing eight or more wavelengths.

The equipment development that made dense WDM technology practical was the optical-fiber amplifier operating at 1550 nm. The combination of dense multiplexing and optical-amplifier technologies provided the capability to transport numerous wavelengths (128 have been demonstrated) down one fiber, amplify them and, therefore, achieve a marked increase in network capacity.

By implementing these two technologies, the BTN can match transmission demand as needed (see Fig. 2). In 2.5-Gbit/sec increments, network capacity growth can be gracefully expanded. On the other hand, in a 10-Gbit/sec time-division multiplexed (TDM) network, excess capa city has to be paid up- front--that is, at network installation. Moreover, because 10-Gbit/sec TDM networking is considered a relatively immature technology, present costs are higher than those expected when this technology matures. In contrast, 2.5-Gbit/sec technology has already achieved a level of maturity.

In developing dense WDM, researchers made the technology backward-compatible; dense multiplexing can be applied to existing networks, thereby accommodating the embedded fiber-optic plant.

Some embedded fiber-optic cable (depending on the manufacturer) experiences high dispersion or optical light losses in handling high-speed signals when using certain multiplexing schemes. With dense multiplexing, however, each wavelength operates at 2.5 Gbits/sec rather than at 10 Gbits/sec. This characteristic allows the network to function reliably despite some of the dispersion limitations imposed by previously installed fiber.

Network design

Representing an economic and flexible hedge on long-term network expansion plans, dense WDM technology allows the capital expenditures for equipment to more nearly match the demands for service. For example, individual wavelength operation can be added to network operation at any time. Network providers can install terminal equipment to handle a few wavelengths now and, perhaps in two years, install the equipment needed to carry more wavelengths. This method can be used incrementally until network capacity is reached.

In this manner, network providers buy the terminal equipment as needed rather than initially overinvesting in terminal equipment when only a small uptick in capacity is required. However, as with all communications networks, there are economic tradeoffs.

Generally, dense multiplexing proves cost-effective on long-distance, high-capacity fiber trunks. When a network route reaches 400 to 500 km, the price of dense WDM electronics is approximately half that of TDM single-fiber systems. However, in certain networks, dense WDM technology also becomes cost-effective on low-capacity, short routes.

As for signal quality, for a given fiber capacity, dense WDM networks can tolerate higher levels of dispersion and polarization-mode dispersion. Studies show that a 10-Gbit/sec TDM terminal is 16 times more sensitive to chromatic dispersion than a 2.5-Gbit/sec system.

In dense WDM networks, though, some overhead expenses are higher than those of TDM networks at installation time because of the cost of the optical line system multiplexing equipment at the network ends. These expenses are incurred whether the traffic rate is 2.5 Gbits/sec or faster. Over time, however, these networks provide more flexibility in the adding or dropping of individual wavelengths. For example, when a 400-km route is initially built, the transmission objective is generally limited to getting information from one location to another. Later, increased communications needs might call for the installation of additional equipment locations between the end points.

With dense multiplexing, the OLS end terminals can be connected to the fibers at the new locations to separate the combined eight wavelengths into individual wavelengths and allow a signal to be dropped out of the network (see Fig. 3). The seven remaining wavelengths would be patched through to their destinations.

New BTN installations can use the nonzero-dispersion Truewave fiber specifically designed by Lucent Technologies` Bell Laboratories for carrying optical signals at multiple wavelengths. This fiber carries the same 1550-nm signals used with conventional dispersion-unshifted fiber but carries them with less distortion.

This characteristic also allows the signals to travel greater distances (2.5 Gbits/sec to distances of 1000 km, and 10 Gbits/sec to 300 km) before requiring amplification.

Optical amplifier vs. regeneration

As light pulses travel through optical fiber, they become weaker, lose their shape and require boosting. Electro-optical regenerators convert optical signals to electronic signals, amplify the signals and then convert the electronic signals back to optical signals before passing the traffic toward its final destination.

On the other hand, optical amplifier repeaters employ special erbium-doped fibers and laser pumps to boost lightwave signals directly without converting them to electronic signals and then back to lightwave signals.

Consequently, optical amplifiers offer several network advantages. Distances between optical amplifiers are much longer than distances between regenerators on nonoptically amplified systems. Longer distances translate into fewer regenerators and lower cost.

Another attribute of optical amplifiers is transparency to the signal transmission bit rate and the modulation format. In dense WDM networks, this transparency enables wavelengths to be incrementally added without replacing the amplifiers in the outside fiber plant. Consequently, higher network reliability is achieved because optical amplification involves fewer components than do electronic regeneration schemes.

End terminals

In the signal transmit direction, the OLS end terminal optically combines eight different wavelengths (dropside signals), amplifies the signals and launches them onto the transmission fiber. In addition, it couples low-level kilohertz-frequency electrical tones on each of the eight wavelengths. These tones are used to monitor the operational status of the eight channels. Even shifts in performance, as well as failures, are observable. Trends in the signal tones can therefore be used to predict failures. In this manner, appropriate repair measures can be undertaken before a breakdown actually occurs.

In addition, the eight-wavelength BTN employs an integrated operations maintenance channel, or ninth wavelength. This wavelength is specifically used to transport a 155-Mbit/sec supervisory signal at 1532 nm.

This telemetry or supervisory signal serves mainly as an operational status check of the optical-amplifier repeaters. It passes measurement information among the repeaters and the end terminals for maintenance purposes. Typical measurements include the optical power of the signals being amplified and the signal-to-noise ratios. Central office craft personnel, therefore, have access to comprehensive maintenance information from the end terminals and the optical repeater sites.

The BTN maintenance goal is to detect failures, monitor facility performance degradation, isolate faults to specific circuit packs, report to an operations system and raise alarm indicators. The supervisory signal continuously monitors the operation of circuit packs and incoming signals. When a fault is detected, the maintenance system employs automatic diagnostics to isolate the failed pack or signal. Failures are reported to a local operations system for repair procedures. All fault conditions can be retrieved through a craft interface terminal. u

Ove Parmlind is global lightwave Synchronous Digital Hierarchy product manager for Lucent Technologies, Merrimack Valley, MA.