By John Dunne
It is no longer easy to decouple component and system functions or vendors. This blurring creates a new layer of responsibility—and opportunity—within the overall delivery chain of optical–fiber telecommunications systems.
The emergence of tunable components is causing significant changes in the relationship between component and system vendors within the optical telecommunications industry. The traditional areas of responsibility of each party are becoming less well defined and the traditional definitions of what constitutes a "subsystem" are broadening.
The first change is a shift in the recognized interface layer between component and system manufacturers. This means that new interface definitions must be designed at a higher position in the traditional supply chain—which blurs the formerly transparent areas of responsibility of both the system designer and the component manufacturer. A clear example of this is the source laser for DWDM systems in which upward movement of the design interface between component manufacturer and system vendor has taken place, landing in a less well–defined region (see Fig. 1).
Traditionally, a DFB single–wavelength source has been supplied in a 14–pin butterfly package with an interface definition provided to the system designer at the level of input electric currents and voltages to drive and stabilize the device. However, even with this single frequency source, the change toward a multisource agreement (MSA) at the transmitter module level has already taken place. This trend is evident in the establishment of the 24–pin MSA for 2.5 Gbit/s single–frequency transmitters between Lucent Technologies, Alcatel, and many of the incumbent source laser providers.
As the number of separate channels in a DWDM system has grown from 4 to 80 wavelengths, tunable laser sources are being developed for sparing and inventory reduction. The tunable laser, however, is considerably more complex to drive correctly and to stabilize to the same standards as the DFB. Therefore, a new interface definition is being sought, primarily through MSAs and the Optical Internetworking Forum. This new interface will require that the system designer provide a DWDM channel number and electrical drive power to the laser. For the basic source market, the job of driving the laser and stabilizing it is the responsibility of the tunable–laser manufacturer.
So far, many tunable–laser manufacturers are recognizing this change in interface definition and a number of leading vendors are making the required adjustment in pursuit of the initial sparing and inventory market. The "DFB replacement" example is probably one of the easier cases to discuss, but there are other markets on the horizon. Consider what happens, however, when more–complex interfaces are required for the next set of applications of the tunable laser.
Likely follow–on applications for tunable lasers are wavelength provisioning and the restoration of transmission links in which the tunable function is used dynamically in the network for fault protection. Routing and switching of optical packets will be the final market to appear as the next generation of systems are designed. Each application will require a new set of interface definitions for the tunable laser. In specific cases, added functionality will be required, thus broadening the range of possible required interfaces even further.
The required skill set to produce the tools and the products will now comprise some capabilities from the component world and some capabilities from the systems world. For example, a faster channel switching speed (for example, 200 ns) would require a faster interface (parallel bus). An additional blocking function (such as an optical switch) may be required during switching events to block noise interference with channels not involved in a switching event. This interface and functional change results in a different interface requirement, which is application dependent. The conclusion is that it is no longer clear who takes responsibility for this new interface.
Some system designers would argue that a common interface for all tunable lasers for all applications should be possible. Possible? Maybe. But practical? The final tunable–laser module might possess too large a form factor and be operationally inefficient due to the redundancy of some of the functionality in certain applications. For example, there may be no need for a parallel interface where switching speed is on the order of milliseconds. Then there is the power consumption and cost of the module to consider. This bloating of features will counteract the system designers' streamlining efforts as they are trying to maintain competitive advantage. Certain applications will therefore require tailored solutions to meet size, cost, and power–consumption targets.
The second change is in the level of integration of previously distinct optical functions. Singly packaged components are now being integrated together to form a more functional final package (see Fig. 2).
Let's take the laser source once again as an example and add to that the functions of modulation and amplification. Previously the system designer purchased a laser source, a modulator, and an optical amplifier as discrete components, and created a subassembly or transmission subsystem by packaging all three. The system designer could optimize the interplay of these components for the benefit of the system.
Now the demand is for an integrated source, modulator, and amplifier, which results in a smaller form factor and a cheaper solution. Who takes on the task of optimization of the functional interaction of these components for specific applications?
In effect, some of the traditional responsibilities of the systems engineer, such as modulation and amplification are being combined with lower–level functions such as laser control. When tunability is introduced into the equation, there appears to be scope for a new type of expertise to emerge. The final module is so functionally complex, that innovative system design will require tailoring of the subsystems on an application–by–application basis.
SUBSYSTEM EXPERTISESo if we consider the two trends—the vertical shifting of interface definitions and the greater level of functional integration—it is clear that they are being caused by the increased demand for active and passive photonic components, especially tunables. These tunable devices are delivering new capabilities for systems that take advantage of the optical wavelength domain.
One characteristic of these new capabilities will be the link between the network management system down to the tunable components. In other words, the traditional multilayered system architecture now suddenly has a thread of control from the uppermost layers down to the physical tunable components. Where will this lead?
This newly created link between management layer and physical layer represents an opportunity for innovation—one that should be taken advantage of by designers of next–generation systems (see Fig. 3). It creates a need for expert subsystem teams that have greater knowledge of the underlying physical layer than before. It also creates a need for a new level of test, certification, calibration, and reporting—elements that are badly lagging in development terms for today's systems (by an order of magnitude).
In fact, this link is a somewhat new phenomenon to beset the old system–component divide, where they have previously existed in a degree of isolation from each other. This opens opportunities for solutions–based companies that focus on understanding how best to take advantage of the link from management layer to the tunable components in the physical layer. These solution sets must also contain the practical solutions for testing components, certifying that the systems using them are working correctly, calibrating (sometimes on an application–by–application basis), and reporting on the more–complex elements while in operation.
John Dunne is founder and CEO of Intune Technologies, 9c Beckett Way, Dublin 12, Ireland. He can be reached at john.dunne@intune–technologies.com.