Optical add/drop multiplexers match network growth

Wdm94067 26

The ability to add and drop wavelengths grows in importance as the complexity and density of WDM systems increase. A typical deployable OADM can now demultiplex eight wavelengths; by 2008 a typical design must support 128 wavelengths. The increasing demands are mirrored in a rapidly growing market for modules that should accelerate from $17.2 million in 1998 to $6.66 billion by 2008.

Jeff D. Montgomery

The emergence of the optical add/drop multiplexer (OADM) is directly tied to the rapid advancement and use of dense wavelength-division multiplexing (DWDM). Earlier, there was limited deployment of optical drop and add at intervals along main trunk lines, but these related to single-wavelength transport fibers; no optical multiplexing was involved, and no identifiable product resulted from this function. The OADM unit, in contrast, involves demultiplexing a number of wavelengths from the total wavelength count of a fiber, dropping these to local subscribers and in turn adding wavelengths from subscribers, multiplexed back into the single fiber output.

The early OADMs, introduced and shipped in 1997 and continuing through 1998, were relatively simple units, involving few wavelengths and few optical switching options. These units were produced by traditional passive fiberoptic component vendors, supplied to equipment vendors for incorporation into their units and supplied directly to carriers. The base price of these units reflected the cost of the demultiplexing filter module and the passive wavelength combiner-generally in the range of $1000-$2000 per wavelength. However, many customers specified special arrangements of wavelength drop choices, and this customizing effort more than doubled the OADM price.


OADM units deployed through mid-1999 were almost all fixed add/drop, but design has rapidly become more complex. The total wavelength throughput of OADM units has advanced from the original four wavelengths to 16 wavelengths, and this rate of wavelength count expansion will continue. DWDM systems with 32 and 40 wavelengths are now common, and systems now contracted for deployment by mid-2000 include units accommodating 64, 80, and 128 wavelengths.

It is broadly expected by carriers that 256-wavelength mux/demux units and related DWDM systems will be introduced by vendors by mid-2000 and will be deployed in active networks before the end of 2001. Looking to the 2003-2008 span, longer-range planning by carriers is focusing on a 400-nm spectrum-approximately 1260-1660 nm-with 50-GHz wavelength spacings or closer, accommodating a nominal 1024 wavelengths. Most interchange (long-haul) carriers expect that by the time this product capability is available for deployment, the bandwidth demand by customers will make the deployment of these systems economically attractive.


A typical deployable 1998-1999 OADM demultiplexes eight wavelengths from the transmission fiber and sends half of them directly through the OADM to be recombined on the output fiber (see Fig. 1). The other four are dropped, two of them going directly to local-subscriber fiber connections, the other two going to receivers for conversion to electronic signals for further demultiplexing in the separate electronic add/drop multiplexer.

At the add end, two fibers on fixed wavelengths come direct from subscribers to the combiner, and the other two wavelengths are originated by fixed precise-wavelength transmitters driven from the electronic add/drop multiplexer output. At the data rate of 2.5 Gbit/s, this provides a total throughput of the OADM of 20 Gbit/s. This unit has no reconfiguration or wavelength-conversion capability.

In contrast, the typical OADM projected for early deployment in 2008 assumes 128 dense WDM wavelengths on the incoming fiber, with full reconfigurability for dropping any selection of the wavelengths from none through all, with all wavelengths that are not dropped being, therefore, passed through (see Fig. 2). The drop-versus-pass-through connections are reconfigurable within microseconds.

At the add end, precise switchable wavelength-locked transmitters are used to recombine up to the total 128 original wavelengths or any portion of those, with any remaining channels converted to wavelengths other than any of the original incoming wavelengths. This model still provides for a fixed share of the drop wavelengths to go direct to optical-connected subscribers, with the remainder going to receivers for conversion and hand-off to the electronic add/drop multiplexer, and with the reverse arrangement at the add end.

This advanced OADM model will operate at 10 Gbit/s per channel, yielding throughput of 1.28 Tbit/s, a growth of more than 60 times over the 1998-1999 model. Currently, research and development is underway on dense WDM capability with compatible optical fiber amplifiers for 256 wavelengths; 40-Gbit/s transmitters will be in production by 2001-2002 (with current research on transmission at data rates higher than 40 Gbit/s). This more-extreme model would receive more than 10-Tbit/s throughput per fiber.


Over the past several years, there has been a continuing movement toward development and deployment of an all-optical network (AON). The AON concept is a transparent network covering an entire continent without any electronic regeneration of signals required enroute. One element of the AON will be rapid reconfiguration capability for disaster recovery, load balancing, and other options, which will require high-performance transparent optical switching, both at network nodes and in the OADM units.

Future OADMs will evolve to take a larger share of deployed units, accommodating more wavelengths. Many of them will include 1 × 2 and/or 2 × 2 optical switches for semi-reconfigurable and fully reconfigurable add/ drop functions. The OADM units will incorporate optical crossconnect matrix switches and other special configurations that can select any or all wavelengths for drop and can combine these, including translation to new wavelengths at the output. So, while the average price of a specific OADM will drop substantially over the forecast period, this countering movement toward a higher share of much more complex and expensive units will keep overall average prices relatively constant.


Although OADMs of 1997 through early 1999 were relatively simple passive units provided mainly by component vendors, many OADMs in future years will be quite complex and expensive, including numerous high-performance components plus operational software. In effect, many of these units will be systems rather than components and are of major interest to traditional optical-network equipment vendors. Equipment vendors, in addition to cultivating their relationships with fiberoptic component vendors, in many cases have independent OADM and related component developments proceeding in their laboratories to support their long-term system needs.

The total OADM function will include one or more optical amplifiers, plus a receiver for each wavelength dropped and a transmitter/transponder for each wavelength added. In some instances, these optoelectronic components will be combined in the overall OADM package; in other situations they will be procured and installed separately from the OADM. Optical add/drop multiplexers are following the deployment trend of dense WDM links, which, in turn, leveraged from the successful introduction of optical fiber amplifiers.

The global consumption value of OADMs will accelerate from only $17.2 million in 1998 to $687 million in 2003, then expand further to $6.66 billion by 2008 (see Fig. 3). This is the forecasted value of the assembled hardware modules, excluding the value of operating software, remote control and instrumentation subsystems, and other system-level costs.


OADMs, as defined for the this study, incorporate a variety of optoelectronic and passive optical components (see "OADM defined," p. 6). As the capacity and complexity of OADMs greatly expand over the next decade, and as the number deployed rises, the related component value correspondingly climbs.

Components include demultiplexing filter modules, passive optical combiners, optical switches, and optical amplifiers, receivers, and transponders, where applicable. These components constitute most of OADM value. The balance is assembly/test labor, related sales/administrative/overhead costs, and profit.

The global consumption of OADM components in 1998 is estimated to be $9.5 million. This figure will rise rapidly to $534 million in 2003 (see Fig. 4). Strong growth will then continue, averaging 59% per year, to reach $5.35 billion by 2008. Components, according to this report, are priced at forecasted market prices, whether purchased or produced internally.

Jeff D. Montgomery is founder and chairman of ElectroniCast Corporation, a consultancy located at 800 South Claremont, Ste. 105, San Mateo, CA 94402. For more information, contact him at 650-343-1398 or at electronicast@msn.com; www.electronicast.com.Wdm94067 26

FIGURE 1. Current OADMs typically demultiplex eight wavelengths from the transmission fiber: four go directly on the output fiber, and four are dropped. At the add end, two fibers on fixed wavelengths come direct to the combiner and two from the electronic add/drop multiplexer.Wdm94067 27

FIGURE 2. By 2008, a typical OADM may have 128 dense wavelengths on the incoming fiber, with full reconfigurability for dropping any selection of wavelengths. The drop-versus-pass-through connections are reconfigurable in microseconds.Wdm94067 28

FIGURE 3. The global consumption value of OADM hardware module-excluding the value of operating software, remote control and instrumentation subsystems, and other system-level costs-rises rapidly to meet DWDM system demands.Wdm94067 29

FIGURE 4. As OADMs play an larger role in optical networks, the value of related components should rapidly rise.

More in Home