4.9 DWDM SYSTEMS BY NETWORK LAYERIn Section 4.3, we described the various layers of the optical communications network and the various DWDM systems in each layer. In this section, we focus on each individual DWDM system, their characteristics and complexities. In particular, we describe point-to-point systems (long-, medium-, and short-haul); large cross-connect systems, optical drop and add multiplexers (single, multiple, and dynamic wavelengths); metropolitan or Metro systems (large, medium, and small), and access systems and first/last mile access. 4.9.1 Point-to-Point Systems 4.9.1.1 Long-Haul Optical Systems Point-to-point (PtP) long-haul (LH) optical systems are designed to interconnect two distant megalopolises (intercontinental) (see upper right part of Fig. 4.5) or continents (transoceanic) and to handle a huge amount of aggregated traffic; such PtP systems are also known as ultra-long haul (ULH) or ultra long-reach (ULR). ULH systems are already in deployment in transoceanic applications, and the globe has already been wrapped around several times (Fig. 4.45). The desired characteristics of a PtP-LH system are to: - Support cost-effectively and reliably a large number of optical channels (wavelengths) per fiber (80, or 160, projected to 1,000)
- Support one of the highest bit rates per optical channel (10 Gb/s, or 40 Gb/s, projected to 160 Gb/s) and RZ or NRZ modulation (and perhaps solitons)
- Support protocol transparency (transport simultaneously SONET, IP, ATM, GbE, and others), and
- Transport traffic over a distance between end points in excess of 4,000 km
 Figure 4.45 An illustrative example of trans-oceanic fiber wrapping the globe and interconnecting every country (an actual picture would have many more fibers).  Figure 4.46 All optical amplification (EDFA + Raman) and dispersion compensation modules (DCM) enable the optical signal to reach ultra long distances (~4,000 km) between end terminals. The optical path between end points is optimized for the minimum possible number of optical amplifiers and optical signal conditioners (chromatic and polarization dispersion compensation), and transponder systems (3R-OEO) are only at each end of the optical path (Fig. 4.46). The key functions of a transponder system are reshaping, retiming, and restoring (amplification), hence 3R. Currently, 3R transponders convert the optical signal to an electrical signal and back to an optical signal (OEO); as such, they are also known as optically opaque. When the signal is converted to electrical, opaque 3R transponders will: - "clean up" the signal from noise (cumulative ASE noise from OFAs), wander and jitter.
- restore its amplitude from fiber loss, component insertion loss and nonlinear effects.
- restore its pulse spectral shape induced from fiber chromatic and polarization dispersion.
In addition, while the signal is in the electrical regime, OEOs perform signal performance monitoring and wavelength conversion/interchange, and perhaps some traffic shedding or grooming, aggregation, and desegregation. However, current 3R-OEO implementations are at a cost premium and therefore cost-effective all-optical 3R solutions are preferred. To meet the required LH characteristics, several ramification methods are used for optical amplification (SOA, EDFA, and Raman), dispersion (chromatic and polarization) compensation, channel equalization, and noise filtering. The design aspects of LH systems comply with ITU-T standards (e.g., G.691) that provide parameters and values for optical interfaces of single-channel long-haul systems. Some of the key components at the outputs of LH systems are: - high quality, high reliability and high-stability lasers (over short and long time) with ultrafast modulators and excellent isolation.
- optical amplifiers to boost the DWDM signal for transmission on the optical fiber.
- EDFAs and Raman copropagating pumps.
- multiplexers combine wavelengths for transmission on a common optical fiber. They also combine supervisory channels.
- optical monitors to monitor the parameters of the outgoing signal.
- modulators.
and at the inputs: - Preamplifiers and Raman counterpropagating pumps.
- Dispersion compensating modules to provide bulk compensation. Compensation requires knowledge of the types of fibers on the path. Ideally, a high-quality fiber with low dispersion and low PMD (<0.1 ps/
 ) is desirable. Nevertheless, because in reality the optical path may consist of different fiber types a common sense margin (~1 to 2 dB) is incorporated in the calculations; this also depends on modulation method, span length, number of spans, line amplifiers and the overall tolerance on differential group delay. - Demultiplexers that separate the DWDM signal into individual channels, including the supervisory channels.
- Highly sensitive and very low-noise receivers and demodulators with stable and adaptive thresholds. This also includes a cross talk margin (~0.5 dB). In addition, depending on bit rate, the optical signal to noise ratio needs to be better than 20 dB and the BER better than 10–16.
- Optical monitors to monitor the parameters of the incoming signal.
Because of the huge aggregated bandwidth that each PtP-LH system handles over a single fiber, high reliability and service protection are required. The protection strategy may be one of three: 1 + 1, 1:1, or 1:N. The 1 + 1 protection strategy requires two identical but separate paths that connect the two end terminals, and each end-terminal feeds the same traffic in the two separate fiber paths (Fig. 4.47). At the other end point, the system monitors the integrity of the received signals on both incoming traffics and selects the traffic from the path with the best performance. This implies that these systems have redundant fabrics, and they have fast monitors to reliably compare the health of one path with another. Clearly, this protection strategy relies on the healthy operation of the receiving node, but it commits resources. However, it does not rely on supervisory messaging between two nodes. It selects a path autonomously and thus it is the fastest; therefore, it is suitable to highly reliable long-haul transmission.  Figure 4.47 To increase service protection, traffic is fed simultaneously in two separate fiber paths and the receiving node selects the best path; this is known as 1 + 1 protection.  Figure 4.48 High priority traffic is transported over the service path and low priority over the protection path. If the service path is at fault, high priority traffic switched onto protection and low priority is dropped; this is known as 1:1 protection. 1:1 protection requires fast and reliable supervisory channel for signaling. The 1:1 is another protection strategy (Fig. 4.48). Connectivity between two nodes is established over a single path (the servicing path), and although a second path (the protection path) is committed, data do not flow over it; in fact, the protection path also may be used to transport low-priority data. In the 1:1 case, the integrity of the signal is monitored and performance messages are sent between the two connected nodes over the supervisory channel. If the performance of the servicing path is degraded below an acceptable threshold level, then the sending node sends messages to the receiving node to switch to the latter path and feeds data into the protection path; if the protection path was transporting low-priority data, then they are dropped. This strategy is slower than the 1 + 1, does not commit the same resources, and is more suitable to medium -and short-haul transmission. The 1:N protection strategy is similar to 1:1 with the exception that for N servicing paths there is only one path committed. In this case, when one of the N paths experiences severe degradation, its traffic is passed over the protection path. Clearly, this case assumes that no two paths will be severely degraded. Finally, in PtP-LH transmission, optical add-drop is limited to a small percentage of traffic, and perhaps using dedicated wavelengths. In this case, the end point systems are provisioned to put add-drop traffic on wavelengths that will be dropped and added by intermediate OADMs. 4.9.1.2 Medium-Haul and Short-Haul Optical Systems Medium-haul DWDM systems, also known as intermediate reach (IR) systems, are very similar to LH systems. However, because the path length is much shorter than 4,000 km (1,000-1,500 km), system requirements and certain design parameters are relaxed to fit the particular business model. For example, they are engineered for more add-drops than the LH, the bit rate per optical channel can be the highest possible (shorter spans can accommodate higher bit rates), and optical amplification may be less cumbersome. In addition, the protection strategy can be 1:1 instead of 1 + 1 (this requires half the network resources but complicates the protection protocol). Short-haul systems, also known as short reach (SR) systems and very short reach (VSR) systems, are much simpler than medium-haul systems because the path distance is on the order of 500 km (SR) or shorter (VSR). They have several add-drops on the path; the total aggregate bandwidth is large; and, like the MH, they require advanced wavelength and bandwidth management. Standards have been developed specifically for this case such as, ITU-T G.691 Short Haul S-64.2 b and Telcordia GR-1377-CORE Intermediate Reach IR-2. In general, manufacturers design systems that can be flexible and growable. That is, they offer platforms that can be customized to meet the needs of specific applications; starting with a minimal system configuration, grow as needed by adding more units (circuit packs) in a shelf, add more shelves in a bay, and add more bays; use input and output ports (transceivers) that fit the specific application; use custom-suited software for provisioning, protocols and management, and so on. Thus, based on a generic platform, the specific application determines the final system complexity. 4.9.1.3 Linear Optical Add-Drop Cross-Connecting Systems Long-haul or medium/short-haul nodes become more cost-effective when they are able to add-drop traffic. Add-drop nodes, also known as add-drop multiplexers (ADM), are known in DWDM systems as optical add-drop multiplexer (OADM) nodes. In long haul, it is not unusual to have several OADMs in a concatenated (or linear) configuration whereby each OADM drops and adds traffic to or from a local network (Fig. 4.49). This is known as "point-to-point with linear OADM" topology. |
