Time-to-market pressures will drive further integration, including active and passive technology.
Ed Miskovic
ExceLight Communications
What defines an optical component? The answer depends on whom you ask. To the component vendor, it is often a discrete device such as a transceiver, a coupler, or an optical filter. To the original equipment manufacturer (OEM) or system integrator, the component may well be a high-level assembly such as a DWDM or an add/drop multiplexer.
Increasingly, the OEM wants to source what the component vendor often terms a "value-added assembly." OEMs want the component vendor to do the integration of active and passive components into a plug-and-play subassembly. They want complex multifunction modules in which the integration is complete and the module ready to go into equipment. OEMs see these value-added assemblies as components for their systems.
The reason for this shift in components is fairly straightforward and stems from two trends. First, the optical-networking market is hot, with a breathtaking pace in the introduction of new equipment and new capabilities. Time-to-market is key, and OEMs want quick solutions from vendors. Many optical OEMs today are startups, under pressure to bring new products to market rapidly to satisfy investors, to secure new rounds of financing, or to achieve early market share. In response, even large, established companies are forced to speed product introductions so as not to be left behind competitively.Second, OEMs want to concentrate on core competencies-adding their own value to equipment. Often, these core competencies involve software, manageability, and other "high-level" system issues. OEMs do not want to become integrators at the component level. They do not have the time, resources, or inclination to do so. What's more, the time-to-market competitive pressures force the make-or-buy decision to become more clear cut-time is on the side of those who source higher-level components and modules rather than committing valuable time and talent to developing them in-house.
Of course, the natural question is, do value-added components lower the module costs? The modules often cost more than the sum of the individual components-you pay for the integration, after all. But the overall applied cost is often lower. You save in board real estate, enabling you to incorporate more functionality onto each card. Artful integration can significantly reduce the per-channel board real-estate costs. And finally, savings are realized in the overall lower costs of the system design, fabrication, and testing.
Is there a trend in where value-added components are going? As successive generations of integration occur, the demand from OEMs will be toward modules that add enhanced network management to make the job of monitoring and control easier. In other words, the trend is toward intelligent, controllable modules, incorporating either local intelligence and control or that derived from the overall network-management system.
Examples of modules that are ripe for such integration are optical add/drop multiplexers (OADMs), coarse and dense wavelength-division multiplexers (CWDMs/DWDMs), optical-fiber amplifiers (OFAs), multipump modules, optical crossconnects (OXCs), and tunable ITU lasers. In short, all the basic building blocks of an optical network are candidates for value-added integration at the component level.
Figure 1 shows a network diagram for a representative WDM system. Notice at this higher view, the network is largely constructed of the modules just mentioned. To show the possible evolution of a value-added, highly integrated component, consider the OADM. The same sort of evolutionary growth is possible with other components; indeed, some trends are common to several modules.
The first-generation OADMs were, and are, fixed-wavelength devices using either three-port optical filters or fiber Bragg gratings with circulators for wavelength selection (Figure 2). The structure is relatively simple and requires no management or control. Wavelengths are dropped or added in a fixed manner, based on these built-in filter elements. Costs are low, but capabilities are rudimentary. The device is built entirely from passive components. Inventory, however, can be cumbersome and expensive if multiple configurations for various wavelengths must be maintained.
As optical networks evolve, an obvious drawback is that the fixed OADM suffers from the inability to select wavelengths. What you buy is what you get, making upgrades or changes impossible without additional modules, component substitution, or possibly a complete module replacement. A tunable OADM overcomes this problem and is achieved by the addition of some level of active components for tuning the fiber Bragg gratings, for example, and the wavelength selection of the "add" channel (Figure 3). You can now configure the OADM to add and drop different wavelengths.
With the addition of a remotely programmable control system, the device is now provisionable for the metro market that may require contiguous add/drop channels over a limited tuning range.
Adding optical switches gives the OADM the capability of selecting channels individually. Any "pre-programmed" wavelength can be selected as a through channel or an add/drop channel. This approach supports higher channel counts but requires network management to operate the switches and thereby select wavelengths.
It is with this sophisticated network management and other software-based capabilities that OEMs increasingly differentiate their products. As value-added components evolve, a key consideration is manageability, which expands the flexibility and, hence, value.
At each step of the evolution of the OADM, active and passive components are being combined to increase the flexibility and configurability of the module. A next logical step is to add tunable lasers to the add channel. Coupled with tunable filter technology or wavelength-selectable OXCs, this added functionality provides enhanced versatility for such operations as dynamically provisioning services for "wavelength-on-demand" applications.
Still missing from this product evolution is any type of closed-loop power control to monitor and adjust power levels for channel balancing. Incoming wavelengths can and will have different power levels based on a number of optical-system influences in the various multiplexer, optical amplifier, add/drop modules, and other inline optical components, as well as in the fiber itself. Using a local closed-loop optical power monitoring and control system, the measured power of the drop channel can be used to adjust the coupled power on the add channels. Figure 4 shows an example of such a module.
The optical power of the dropped channels is continuously measured by the on-board monitoring system. A corresponding control signal is used to adjust the power/inline attenuation of the add channel to ensure that the optical power level of the added channel that the network receives is the same as that of the dropped channel (also accounting for insertion losses of the module). In this manner, there is a minimum of channel imbalance for the other system-level components to compensate for.
A logical progression of technologies for future OADMs follows the normal course of technology developments. These include high-speed optical mesh switching, integrated automatic power adjustments, tunable lasers, and improved high-performance semiconductor optical amplifiers (SOAs) for active-channel balancing.
A tremendous amount of activity is ongoing for the development of tunable lasers. Much has been written about the array of technologies used for tuning these various laser designs (temperature-tuned distributed feedback, distributed Bragg reflector, vertical-cavity surface-emitting lasers, etc.). It is anticipated that a major application for tunable lasers, particularly in the metro market, will be in dynamically provisionable OADMs.
A gating item for implementation of these tunable lasers in metro OADMs will be cost. Just as the metro market is anticipating the advent of low-cost optical amplifiers using uncooled pump lasers to fulfill its quest for major system proliferation, it is expected that the metro market will also take advantage of low-cost tunable lasers to further enhance the metro system capabilities and deployment.
It is expected that initial system designs using tunable lasers will require a small tuning range (i.e., 2 to 6 nm). More complex systems with enhanced provisioning will require a much wider range of wavelength tuning (i.e., 15 to 20 nm, or even more). Tunable lasers are now available that will allow tuning over the entire C-band.
As previously noted, the evolution is from simple devices to more complex modules with integrated functionality. As the level of integration and functionality increases, the system-level performance of each module will become more dependent on higher levels of network management. High-speed provisioning will become a necessity, but this provisioning will also include enhanced network management to allow control to be brought down to the level of individual wavelengths.
The evolution from simple components to value-added components runs, as it were, on two parallel tracks. One track is simply to integrate existing technology to add greater sophistication, flexibility, and manageability. The second track is to evolve component technology. It's not only a question of integrating today's components, but of evolving existing components and devising new ones. Ad vances in micro-electromechanical system (MEMS)-based optical-component technology, tunable filters and lasers, and high-performance SOAs, for example, are fundamental to the continual evolution of lower-cost, higher-capability optical communications.
Similar integration will occur to create value-added modules for other parts of the optical communications systems, including OXCs, CWDM/ DWDMs, and OFAs. While the specifics of integration will vary with the application, a common thread is to create components that are manageable, and management requires that such value-added components include both passive and active elements. Efficient network management requires the ability to select wavelengths through tunable components or to switch wavelengths among ports via MEMS-type switches or wavelength routing within the optical-switch fabric.
Dynamic provisioning, remote monitoring and configuration, and other network-management issues depend to a large degree on successful integration of active and passive components into these value-added assemblies. But as the OEM increasingly views such assemblies as components, the distinction of active and passive will diminish and become blurred, immaterial in that capabilities and functionality transcend such simple distinctions.
CWDM will play an integral part in providing low-cost solutions for the metro market. One obvious area of significant cost savings is that of the source lasers themselves. Instead of expensive, cooled 14-pin butterfly ITU lasers, CWDM systems will incorporate uncooled, coaxial package lasers with wider channel spacing. A logical step toward value-added integration in this portion of the metro space would be to incorporate the CWDM transmitters, multiplexer, and channel-balancing attenuators into a single module design.
Eventually, there will be movement to provide some type of standardization for these value-added modules just as that undertaken for the various data-link products. Module customers will want vendor-to-vendor interoperability with well-defined specifications, packaging, connections, etc.
Component technology will continue to evolve, both at the basic and the integrated level. The OEM's demand for integrated, value-added components spurs development. The all-optical network today is a clearly realizable goal, and OEMs are partnering with the component vendors to increase capabilities and lower costs. Eventually, the value-added modules and subsystems will be viewed simply as another component for building better, more efficient, and cost-effective systems.
Ed Miskovic is director of marketing at ExceLight Communications (Durham, NC) and a member of Lightwave's editorial advisory board. The company's Website is www.excelight.com.