Dispersion compensators move onto optical-network fast track

Major drivers of the dispersion compensator market include increased use of dense WDM, higher data rates, use of optical add/drops, and greater distances required before signal regeneration. By 2009, tunable (variable) DCMs will claim a 65% ($283 million) market share for systems with >10 Gbit/s data transmission rates. Stephen Montgomery

Stephen Montgomery

president

ElectroniCast Corp.

With the increased use of dense wavelength-division multiplexing (DWDM), the fibers of network long-haul cables transport a wide range of signal channels and a wide range of data rates (155 Mbit/s to 10 Gbit/s now, 40 Gbit/s by 2002) on numerous wavelengths (32 now common, 128 now commercially available, 256 in 2000) across a wide spectrum. As DWDM puts more wavelengths (channels) onto the same single fiber and transmission rates climb ever higher on those wavelengths, carriers also need to handle dispersion problems. And this is happening at a time when, because of age and multiple vendors, there is a wide variation of characteristics of the various fibers that have been installed over the past 20-25 years.

The characteristics of a fiber affect the nature of a signal pulse as it traverses the fiber. A major concern, especially in long-haul networks, is chromatic dispersion, which is both data-rate and wavelength dependent. (There is increasing concern, also, with polarization dispersion.) At intervals along the fiber, it is necessary to insert dispersion compensators to cancel the pulse dispersion that has occurred. The most convenient points along the trunk to insert these compensators are at optical amplifier nodes, currently spaced 40 to 80 km apart.

At present, most dispersion compensators (DCMs) are installed along a long-haul network link at the optical fiber amplifier (OFA) points to handle the chromatic dispersion problem, thereby lengthening the distance between regeneration points. Regenerators are now priced at several million dollars; therefore, the fewer generators used, the better. Because compensators cause optical power loss, amplification is needed to counter this loss, as well as the node-to-node fiber loss. Long-haul trunk optical fiber amplifiers now comprise two (or more) stages. The most convenient and logical point to insert the compensators, therefore, is between stages of an amplifier.

Dispersion compensation became of concern as deployment of 2.54-Gbit/s (OC-48) channels ramped up, and they become urgent as OC-192 (10 Gbit/s) deployment accelerates, along with the trend to increased spectral width of dense WDM. The growth of the dispersion compensator market from 1997 to 2000 has been explosive (see "Suppliers of dispersion compensators," p. 24).

Combating chromatic dispersion

There are several techniques for minimizing the chromatic dispersion problem; dispersion-compensating fiber (DCF) and chirped fiber gratings are the most accepted methods. DCF has high levels of dispersion of opposite sign to that of the optical signal carrier fiber (standard single-mode fiber). To compensate for the dispersion over an 80-km span of standard optical fiber, a 12 to 16 km length of DCF is linked with the standard fiber in the network. DCF, however, is considered by some vendors and carriers as too large, and it also demonstrates high attenuation and increased optical nonlinear effects.

Fiber Bragg grating-based dispersion-compensating modules are an appropriate solution, because the grating period is chirped to reflect slower wavelengths by the faster ones that must travel further into the module before reflection occurs. As with other fiber Bragg grating devices, an optical circulator is used to segregate the input of the module from the output. As distances and transmission rates increase and DWDM is used with different types of optical fibers, the need to adjust (or tune) the compensating function will increase.

Enter dispersion compensators

There are two common techniques to deal with dispersion in optical networks-the use of dispersion-compensating fiber and the use of dispersion-compensating modules (DCMs) that are typically fiber Bragg grating or other filter-based devices. ElectroniCast`s 1999-2009 forecast for telecommunication fiberoptic dispersion compensators (modules)-specifically grating- and filter-based packaged devices-projects steady increases in market share (see Fig. 1). In 1999, the total global consumption market value of dispersion-compensating devices (both DCF and DCMs) was $324.1 million. DCF represented 89% or $287.5 million in 1999. By 2009, the consumption of DCMs will represent about 74% of the relative market share or $435.5 million, as the need for remotely tunable devices drives the market.

The report also includes a market overview of dispersion compensation by using specially designed optical fiber (see table on p. 24). The North American consumption value of grating dispersion compensators will expand rapidly from $17 million in 1999 to reach $176 million in 2009.

Fiber Bragg gratings perform functions such as stabilization of 980-nm pump lasers by creating feedback to the laser, linear discrimination (when integrated in demodulation units in grating-based sensor systems), DWDM filters (in which fiber Bragg gratings are used to isolate the different wavelengths of the ITU grid), spectral shaping, attenuation, dispersion compensation, EDFA gain flattening, and other functions.

Fiber Bragg grating technology allows for the manufacturing of chirped long-length gratings (>10 m long). This technique allows chirped Fiber Bragg gratings of any length to be made by programming that purpose into a computer-driven command, thus allowing for high-volume production. Previously, the production/manufacturing process was restricting the various fiber Bragg grating inscription parameters.

Because optical add/drop multiplexers (OADMs) require demultiplexing of the wavelengths and later remultiplexing along with amplification, they are a natural point for compensator insertion. Manually inserted compensators are not counted in the as-shipped OADM value, because they typically are added separately, after OADM installation. However, for reasons discussed below, a trend is developing to use remotely adjustable ("tunable") compensators that are built into the OADM, so these devices are counted as part of OADM component value. The OADM consumption, however, will constitute a relatively small share of the total tunable compensator market. Compensators will be needed at every amplifier node (and regenerator), but a relatively minor share of amplifiers will be consolidated with OADMs. Also, the physical location of the OADM needs to be convenient to the subscriber base it serves and this may not coincide with the optimum locations of amplifier nodes.

Generally, a subscriber base (typically, a small/medium population city) can be adequately served by one OADM node, which will incorporate four OADMs-two for bidirectional transport of the active traffic and duplicates for the two as a fiber protection circuit. Typical long-haul trunk cables, however, range from 48 to 144 fibers, and access loop cables are trending toward 864 fibers and up.

Major R&D efforts have been applied to tunable fiberoptic components (compensators, attenuators, laser diodes, photodiodes, and such) over the 1997-99 period, and these components are now emerging into the commercial market. Their major advantage is elimination of the labor (and overhead) cost of dispatching a craft crew to the OADM and amplifier nodes at each of the remote sites each time there is a need to change the compensation. These "truck rolls" are quite expensive, and there is an increasing shortage of skilled craft persons.

Market drivers

According to the ElectroniCast report, fixed modules dealing with telecommunication networks of less than 10 Gbit/s held a 60% relative market share or $22.06 million in 1999. However, by 2004 as the need to remotely adjust higher data rates (>10 Gbit/s), variable DCMs serving 10 Gbit/s and higher links will dramatically increase in market share. By 2009, advanced (>10 Gbit/s) variable DCMs will continue to lead in market share with 65% or $283.11 million. Standard variable modules for network transmission rates of less than 10 Gbit/s will have a 21% market share, or $90.86 million.

Much of the development of gratings processed in optical fiber or in a holographic medium (such as a lithium niobate crystal) has occurred over the past decade and is strongly patent-protected. Licenses, however, are being sold. Some original-equipment manufacturers (OEMs) have adopted this as a key captive element of their DWDM systems. Several other startups are in business or in process.

Eventually, the module market will be led by variable or tunable chromatic dispersion compensator modules (see Fig. 2). Tunability permits users to optimize network link architecture and configuration to accommodate all of the different types of optical fibers and optoelectronics (transmitter/receiver pairs). Variable DCMs will be increasingly important as carriers upgrade their networks to 40-Gbit/s systems. For these systems, especially where long-haul is concerned with several, often difficult-to-reach, remote sites, tunable modules will be in huge demand.

FIGURE 2. By 2009, the global market forecast for types of dispersion-compensating filter modules finds advanced variable devices replacing standard fixed devices for the lion`s share of the telecommunications market.

Suppliers of dispersion compensators

AC Technology Inc.

ADC Telecommunications (including Princeton Optics)

Advanced Optronics Corporation

Amphenol Fiber Optic Products

Avanex Corporation

Bragg Photonics Inc. and QPS Technology Inc.

Corning Incorporated

Corvis Corporation (including Kromafibre)

DiCon Fiberoptics Inc.

E-TEK Dynamics

FOCI - Fiber Optic Communications Inc. (MRV Communications)

HighWave Optical Technologies

Innovative Fibers Inc.

INO

IONAS A/S

JDS Uniphase

Kaifa Technology Inc. (acquired by E-TEK)

Kromafibre (Corvis Corporation)

LaserComm Inc.

LightPath Technologies Inc.

Lucent Technologies

MRV Communications (including FOCI)

New Focus Inc.

Newport Corporation

Nortel Networks

NSG America Inc.

Optics For Research Inc.

OZ Optics Ltd.

Princeton Optics (acquired by ADC)

QPS Technology Inc. (Bragg Photonics)

RoMack Inc.

Samsung Electronics

SDL Inc.

Sumitomo Electric Lightwave Corp.

3M Telecom Systems

Wave Optics Inc.

Source: ElectroniCast Corp.