PDF of this tutorialThe Internet has forever changed the way companies function and the way people live, work, and play. As though fulfilling unwritten laws of physics and economics, bandwidth demand has expanded to constantly “push the edge of the envelope” of network capacity, with useful, novel applications springing up that immediately test the limits of ever-improving networks. These include peer-to-peer computing—most famously, Napster—Webcasting, storage area networks, and the coming growth of video-on-demand that will deliver VHS–quality movies into the home via a broadband Internet connection. Bandwidth-hungry applications are driving usage growth rates to more than 100 percent per year; as a result, many trunks that were lit only 30 percent in 1997 have now reached the point of fiber exhaust.
Investing in the “Network of the Future”
As businesses and individuals continue to exhibit an insatiable demand for bandwidth, the telecommunications industry and its suppliers have become highly motivated to build networks that deliver data farther, and faster, than ever before. Cost-effective bandwidth growth requires the upgrade of existing fiber links—an installed base that’s growing at more than 10 percent per year—in addition to laying new fiber, which carries an installation investment of up to $.5 million per kilometer.
So although 2.5 Gbps OC–48 networks were widely perceived as “ultra high speed” just a few years ago, telecommunications companies are now rapidly ramping up capacity on many links to 10 Gbps OC–192, and the industry has set its sights on 40 Gbps OC–768. Many industry observers believe that OC–768 networks will become a reality within the next 12 to 24 months.
The Need for Tunable, Multichannel Solutions
Managing dispersion in high-speed optical networks is just one component of the overall challenge designers are contemplating as they start to plot OC–768 “networks of the future.”
As the industry moves toward full-mesh, all-optical networks, unprecedented new levels of flexibility and manageability will be within reach. In this OC–768 environment, network traffic will flow over many dozens of channels and be dynamically routed around network faults. Today’s tedious provisioning process, which can take days or weeks, will be replaced by a far more dynamic management environment in which customers themselves can secure a specific bandwidth allocation for a fixed length of time—to handle a global Webcast, for example, peer-to-peer traffic, or myriad other applications.
Figure 2. In the dynamic, highly reconfigurable networks of the future, tunable, multichannel dispersion compensation modules are required to adapt to variable path characteristics.
The easily reconfigured “network of the future”—based on dense wavelength division multiplexing (DWDM) technology and characterized by streamlined channel switching, channel add/drop, as well as dynamic path reconfiguration for bandwidth allocation and restoration—requires its components to be similarly flexible. As it pertains to dispersion compensation modules, specifically, tunability and inherent multichannel capabilities become issues of paramount concern.
Tunability is the ability to optimize the amount of compensation delivered by a dispersion compensation module (DCM) to precisely match network requirements. A DCM can be tuned manually, remotely, or adaptively or through a combination of techniques. Manual tuning is performed by a network technician who adjusts DCMs prior to or after their installation on telecommunications racks. Remote tuning is done from a central console, using network management software. Adaptive tuning, as the term suggests, is a dynamic, intelligent process executed within the DCM, without any human intervention. From a dispersion compensation perspective, adaptive tunability is essential in allowing “networks of the future” to manage change as they adapt to variable path characteristics, environmental fluctuations, and configurations that are themselves in a constant state of change.
Within DCM units, native multichannel capability is similarly essential; it eliminates the current “one channel, one box” ratio required by rudimentary first-generation dispersion compensation devices. In many current solutions, one DCM must be provided for each channel. In DWDM environments with just a few channels, this situation is manageable. However, in complex OC–192 and OC–768 environments, space and cost constraints demand that DCMs have multichannel functionality.
DCMs with multichannel capability are also a vast improvement over the most common solution used today, spooled dispersion compensating fiber (see Topic 3). DCF, as it is commonly called, is a chromatic dispersion solution optimized for a specific wavelength within a network link. Although DCF produces acceptable results in today’s networks, as network fiber becomes upgraded and loaded with more channels, its performance becomes sub-optimal. In addition, DCF only partially addresses slope mismatch and does not compensate for polarization mode dispersion.
