Chromatic dispersion (CD) is a distance-related phenomenon that impedes the quality of an optical signal. In the past, dispersion-compensating fiber (DCF) was the only available method to mitigate its effects. Today, a new class of filter-based variable CD compensation devices has emerged, hoping to tap into a worldwide market that should increase from its $580,000 mark in 1999 to $30 million by 2005, according to a new study from market researcher ElectroniCast (San Mateo, CA).
CD occurs because different frequencies travel at different speeds down the transmission fiber, causing the optical signal encoded on those frequencies to smear and merge together. The receiver cannot distinguish between neighboring bits in the optical data stream, resulting in a high bit-error rate. CD is a critical problem in long-haul as well as metro core and regional networks, particularly as system operators upgrade to 10- and eventually 40-Gbit/sec data rates.
By compensating for CD, it's possible to increase the signal quality and reduce the bit-error rate, thereby increasing the distance capabilities of a given system. Almost all the CD compensation occurring in today's networks is done via DCF. In the typical long-haul system, DCF can be found at amplifier sites.
While it remains the best solution for some applications, DCF has its limitations. Spools of fiber are often needed to compensate for CD in a network, which results in higher insertion loss. Longer lengths of fiber suffer from higher insertion loss, which requires additional amplification. Amplifiers add noise, which requires additional filters and regenerators and so on.
Perhaps the biggest disadvantage of DCF is that it only corrects a fixed amount of dispersion at a given wavelength. Channels at the extreme of the C- and L-bands are often either over- or under-compensated, creating an accumulation of residual dispersion along the fiber. That is where variable-filter-based dispersion compensators will find their market niche.
Higher data rates, an increase in optical-fiber link lengths, and the use of DWDM and optical add/drop multiplexers (OADMs) are driving the demand for filter-based variable dispersion compensators. The passive devices are either Etalon-based or fiber Bragg grating (FBG)-based and typically used in conjunction with OADMs and optical amplifiers.
Variable-filter-based products have an edge over DCF because the dispersion compensation can be remotely controlled or even dynamically controlled, explains Stephen Montgomery, president of ElectroniCast. "It can be adjusted continuously to deal with any problems," he notes. "When you get into higher data rates especially and more mature links, you may need some kind of dynamic control, and that's where the filter-based products come in."
An etalon-based CD compensation device induces periodic time delays specifically designed to reverse the time-delay problems inherent in the transmission fiber. An etalon is best defined as a pair of mirrors facing one another separated by an optical path. Incoming light bounces back and forth between the two mirrors until it exits through one of the two mirrors. The mirrors reflect part of the incoming light and transmit the rest.
For any given etalon, some colors of light will be retained or resonate longer than other colors. The better a given color resonates within the etalon, the longer it stays in the etalon, inducing a time delay. This time delay is a negative of chromatic dispersion created by the transmission fiber.
Accumux Technology (Camarillo, CA) has developed the Light Fixer dispersion compensation module (DCM), based on its patented etalon technology. The Light Fixer DCM enables 80 channels of slope-compensated, tunable dispersion compensation across the C- and L-bands. According to Ian McMichael, principal engineer at Accumux, the DCM device offers low insertion loss, excellent slope compensation, and a more compact package than competing technologies.
Accumux has several tier one customers, all of whom have conducted trials, tested the devices in their labs, and "verified that the devices have the performance that we give on our specification sheets," reports McMichael. "We're just waiting for them to ramp up in volume."
Etalon-based devices provide several key advantages over their FBG-based counterparts, says McMichael, particularly in the area of group delay ripple. "With dispersion compensators," he explains, "you're trying to correct for the kinds of things that naturally happen in a fiber—different wavelengths traveling at different speeds in a fiber. If you're trying to fix that, how well you do determines the maximum length of fiber for which you can compensate. FBGs have a lot of group delay ripple, which is a problem if you want to compensate for long lengths," he says.
The folks at TeraXion (Quebec City), manufacturer of FBG-based devices, disagree. They contend that their ClearSpectrum line of CD compensation products provide ultra-low group delay ripple on the order of <±5 psec.
Historically, FBG CD compensators had been known for their limited bandwidth, which hampered their success in DWDM networks. TeraXion claims it has cleared this hurdle, thanks to advances in holographic phase mask fabrication, a technology derived partly from its December 2002 acquisition of Phaethon Communications. Today, the company offers both static and tunable devices, and the tunables are either single channel or multichannel.
An FBG filter is an in-fiber device; it's written into the core of the fiber via the interference of two ultraviolet beams. The interference pattern forms a periodic refractive index change longitudinally along the fiber. Each index change acts as a series of reflectors. Because there are typically tens of thousands of these index changes in a row, they add coherently; light traveling down the fiber will reflect strongly off the grating structure. FBGs transmit most of the light going through the fiber except for a narrow band where the light is reflected back up the core in the direction of the signal entering the fiber.
TeraXion has been selling and shipping its FBG-based compensators since early 2003 and reports that the ratio between fixed and variable devices is roughly 50:50. Right now, the bulk of the company's variable compensators has been shipped to the government or military for use in secure applications. "We're working with government contractors that require tunable dispersion compensators for 10-Gbit/sec applications, but we don't know more than that," offers Erick Pelletier, vice president of sales and marketing at TeraXion. "There is a market for this; it's not big, but it's real."
Pelletier also reports a resurgence of interest in reconfigurable networking and 40-Gbit/sec data rates. "Some programs were stopped, but now they are restarting," he says. "There's some sampling activity here on our side. We're working closely with customers here to make sure that our solutions will be there in the systems of the future."
So just how big is the market for these filter-based CD compensation devices? Given the state of the long-haul and ultra-long-haul sectors, there's not a lot of demand—yet, says Montgomery. "The whole idea of going up to several hundred channels per long-haul or ultra-long-haul link is not an immediate concern right now," he notes. "The channel counts have come down from what people thought they would be; it's still somewhat manageable with a spool of fiber."
That said, Montgomery believes market demand will ramp as data rates and channel counts increase and as existing dark fiber is lit. Much of the installed fiber was designed to support 2.5 Gbits/sec, not 10 or even 40 Gbits/sec. The longer the market takes to materialize, the more mature the fiber—and the more problems that fiber may have. While there aren't many vendors publicly active in the space, Montgomery believes there are three or four companies that currently have the capability to produce filter-based CD compensation devices and will do so as soon as demand emerges.
Of course, the longer the market takes to materialize, the more competition filter-based variable compensators will face. Electronic and optical chip-based products still lag behind, but "every month this doesn't happen, the chip capabilities get closer and closer," observes Montgomery.