Before any “network of the future” can be considered, the “network of today” must be dealt with. This includes legacy fiber, its ability to handle OC–192 and OC–768 speeds, and the physics of dispersion.

Dispersion Defined

As described in Topic 1, CD, slope dispersion, and PMD cause waveforms within optical fibers to spread out, lose their shape, and become difficult to detect by receivers at the end of a fiber span. Unchecked, dispersion causes bit error rates to increase to unacceptable levels.

• Chromatic dispersion is based on the principal that different-colored pulses of light travel at different speeds. More technically, CD is the sum of material dispersion and waveguide dispersion. “Material dispersion” is caused by the variation in refractive index of the glass in the fiber as a function of the optical frequency. “Waveguide dispersion” is due to changes in the distribution of light between the core and the cladding⁶ of single-mode fiber.

Figure 3. Chromatic dispersion becomes a significant problem at network speeds of OC–192 and higher.

In today’s 2.5 Gbps OC–48 networks, the impact of chromatic dispersion is minimal. Unfortunately, dispersion effects do not increase linearly; the growth is at a rate of the square of the increased speed of the transmission. Therefore, at 10 Gbps (OC–192), CD has a major impact on network performance because it is 16 times worse than at 2.5 Gbps (see Figure 3). Similarly, at 40 Gbps, dispersion levels are 256 times higher than at OC–48.

• Slope mismatch dispersion is a subset of chromatic dispersion. It occurs in single-mode fibers because dispersion varies with wavelength. This can result in a significant buildup of dispersion, especially at the extremes of a band of wavelength channels (see Figure 4). Full compensation of this type of dispersion requires slope matching or tunable compensation at the receiver.

Figure 4. In conventional solutions, slope mismatch dispersion must be addressed individually on each channel of a DWDM network

As indicated in Figure 4, uncorrected “slope” dispersion at channels away from the center channel (λ₂) accumulates over successive transmission links and can reach unacceptable levels.

• Polarization mode dispersion occurs as light travels down single-mode fibers, in two polarization modes. When the core of the fiber is asymmetric, the light traveling along one polarization moves slower or faster than the light traveling along the other polarization. This can cause the pulse to spread enough to make it overlap with others or distort the shape of the pulse enough to make it undetectable by the receiver (see Figure 5).

Figure 5. PMD changes randomly with environmental effects such as temperature and imperfections in fiber cores.

PMD levels rise directly with network speeds (i.e., PMD is four times as severe at 40 Gbps OC–768 as it is at 10 Gbps OC–192) and also increase proportionally to the square root of the distance. For example, if the span length is doubled, PMD increases by a factor of 1.4.

In addition, PMD is a dynamic phenomenon, changing randomly with environmental effects such as temperature and infinitesimal asymmetries in the fiber core (see Figure 6). For this reason, PMD requires adaptive compensation from adaptively tunable DCMs.

Figure 6. PMD levels can range from mild to severe on OC–192 networks, influenced by a range of factors.

Overcoming Dispersion: A Unified Approach

Given the complexity of the dispersion challenge and the fact that its real-world impact is only beginning to be experienced and understood, three principles are emerging that promise to help companies address dispersion globally, throughout the network.

The first principle involves tackling the trio of dispersions—CD, slope mismatch and PMD—on the network level as a group, rather than individually. By taking a global view of the dispersion challenge, companies can best understand the interaction between CD, slope mismatch, and PMD and explore synergies between the various dispersion compensation solutions. This approach also best leverages the time and resources committed to exploring dispersion and helps ensure that chosen solutions will work together seamlessly, in a synergistic manner.

The second overarching principle that can be effectively employed in assessing dispersion compensation solutions is tunability, which allows the chosen DCMs to adapt to changes in the network environment. As discussed in Topic 2, adaptive tunability is essential in enabling 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. Tunability is especially critical for polarization mode dispersion solutions, which must adapt to random environmental fluctuations.

Finally, multichannel capability is critical simply because “one channel, one box” solutions are highly inefficient at OC–192 and OC–768 levels. This is due to physical space constraints on network racks, as well as power consumption considerations.

Current Approaches Fall Short

Although today’s networks have an OC–192 ceiling, several solutions exist that partially address the dispersion challenge. These include the following:

• Non-zero dispersion shifted fiber (NZDSF)—A new approach to manufacturing fiber has produced NZDSF, which is superior to traditional single-mode fiber because it utilizes a different design—a more complex refractive index profile—that results in less dispersion. NZDSF is also manufactured to higher standards, which produce a more perfectly circular core. This creates less dispersion-creating variability in the fiber, thereby allowing signals to be transmitted longer distances. However, dispersion still accumulates with distance but at a lower rate.

Although NZDSF partially addresses PMD at OC–192, it still requires chromatic dispersion correction; in addition, the slope mismatch problem is more severe in NZDSF fiber than in single-mode fiber.

Finally, NZDSF is a limited solution in that, like all fiber, it is costly to install and will only comprise a small portion of the global fiber-optic network for the foreseeable future. As previously noted, the majority of installed fiber is single-mode.

• Dispersion compensating fiber (DCF)—is the traditional solution for handling chromatic dispersion. Spools of DCF are placed at intervals along the network—approximately 15 kilometers of DCF for every 80 kilometers of network fiber—and are typically stacked atop telecommunications racks (see Figure 7).

While each spool of DCF adequately solves chromatic dispersion, if it is used on fiber carrying multiple wavelengths, it is usually set to correct CD most accurately on the center wavelength. However, dispersion still accumulates at other wavelengths and can be a significant problem at the edge of a band of wavelength channels. DCF also does not address PMD.

Figure 7. Chromatic dispersion has traditionally been addressed with spools of dispersion compensating fiber placed at intervals along the network.

Exciting New Developments: DCMs

The newest approach to taming CD, slope mismatch, and PMD is with dispersion compensation modules (devices), a development hotbed within today’s optical networking industry. DCMs are placed in front of receivers on the network and make continual adjustments to the signal, based on information derived from analyzing a sample of the optical pulse as it travels through the module. The degree to which a pulse is corrected is based on its state, read by the DCM’s detector as pulses pass through it.

DCMs offer the most promise in the race to beat dispersion, for several reasons:

• A wide range of tunability options, including remote and adaptive
• Low cost and ease of replacement (relative to alternatives such as laying new fiber)
• Small form factor (several inches, as opposed to cumbersome DCF spools)
• High levels of innovation

Fiber Bragg Gratings: Maximum Efficiency

One of the most advanced technologies being incorporated into DCMs is fiber Bragg gratings, short lengths of optical fiber that reflect a particular wavelength. Fiber Bragg gratings feature periodically spaced zones in the fiber core that have been altered to have different refractive indexes slightly higher than the core. This structure selectively reflects a very narrow range of wavelengths while transmitting others (see Figure 8).

Figure 8. In fiber Bragg gratings, a periodic refractive index variation along the axis of the fiber causes narrowband refraction.

The name comes from Bragg’s Law, which describes the optimal spacing of the changes. Sir William Lawrence Bragg, noted British physicist (1890–1971) made this discovery in his study of X-rays and crystal structures. ⁷

Fiber Bragg gratings are typically between one millimeter and 25 millimeters long. To create the appropriate stack of high- and low-refractive-index regions along a piece of optical fiber, its refractive index must be permanently modified via a photosensitive effect. This is accomplished by exposing the optical fiber to ultraviolet (UV) light.

Compared to most other fiber-optic components, fiber Bragg gratings are simple to manufacture. And unlike micro-optic devices such as thin films and isolators, they do not require complex and precise alignment. Perhaps the biggest advantage of fiber Bragg gratings is their low insertion loss when placed onto the network. This is because the alteration (grating) is made to the fiber itself, allowing the light wave to continuously remain within the fiber.⁸

⁶Cladding is the plastic or glass sheath that is fused to, and surrounds, the core of an optical fiber. Its mirror-like coating keeps the light waves reflected inside the core. The cladding is covered with a protective outer jacket. Definition from Computer Desktop Encyclopedia.
⁷ Definition from Computer Desktop Encyclopedia
⁸ Excerpted from “Fiber Bragg Gratings,” SPIE’s OE Magazine, January 2001