Optical-fiber selection considerations with evolving system technologies

Th Pg62 Table1

Singlemode fiber (SMF) continues to be the dominant fiber type in use worldwide. Various types of premium-priced dispersion-shifted fiber (DSF) have entered the market over the past five years and captured almost 10% of fiber demand. These fibers have traditionally provided enhanced dispersion performance advantages for long-haul routes, although SMF is proving to be more economical and offer superior performance with many emerging high-capacity applications. Let's take a look at changes in system technologies and their economic and technical impacts on optical-fiber selection.

Optical-fiber systems have evolved rapidly over the last few years. New technologies continually appear to address the key system limiting factors in the design of optical networks. These limiting factors include:

  • loss budget or attenuation
  • signal-to-noise ratio (SNR)
  • chromatic dispersion
  • polarization mode dispersion (PMD)
  • nonlinearities due to four-wave mixing (FWM)
  • nonlinearities due to cross phase modulation (XPM)
  • nonlinearities due to stimulated Raman scattering (SRS)
  • nonlinearities due to stimulated Brillouin scattering (SBS)Th Pg62 Table1

    Examples of such technologies are shown in Table 1 in the order of their appearance in the network. These new technologies have enhanced the capabilities of installed optical fibers as well as widened the choices for fiber selection for new routes.

    In the process, this technological evolution has once again made standard singlemode the best economical and technical choice in most installations, including many long-distance applications with new 10-Gbit/sec systems. A review of how the capabilities of SMF and nonzero DSF (NZ-DSF) have changed as system technologies evolve illustrates this evolution:

    • SMF transmission distance capabilities shrank with the introduction of higher-bit-rate systems, providing the impetus to launch DSF, and later, NZ-DSF.
    • Narrow-spectral-width lasers chang ed the preference back to SMF, making it the preferred and most economical choice for 2.5 Gbits/sec and lower.
    • 10-Gbit/sec applications shifted the fiber preference to NZ-DSF for routes with 10-Gbit/sec long-haul requirements.
    • Dense wavelength-division multiplexing (DWDM) forced several NZ-DSF product evolutions to reduce fiber nonlinearities.
    • Dispersion-compensating fiber (DCF) and dual-stage amplifiers permit SMF to achieve the same distances as NZ-DSF at high bit rates, making fiber choice a decision based on economics and future upgrade capabilities by providing superior handling of nonlinearities at higher data rates.

      The current choice for optical fiber includes SMF or various types of NZ-DSF. Standard SMF with the zero-dispersion point located at 1310 nm permits operation in both the 1310- and 1550-nm regions. NZ-DSF with the zero-dispersion point located below the 1550-nm operating region primarily operates in the 1550-nm region.Th Pg64 Table2

      Current distance capabilities for SMF and NZ-DSF fiber are shown in Table 2. Actual distances will vary depending on the number of DWDM channels deployed.Th 07spr061

      Fig. 1. System component requirements for both 2.5 and 10-Gbit/sec systems operating over 600 km of both singlemode fiber and NZ-DSF are provided. The design is identical for 2.5-Gbit/sec systems. 10 Gbit/sec systems operating over 600 km of fiber require more DCF compensation using SMF although end-to-end costs may be less.

      Figure 1 shows the component requirements for 2.5- and 10-Gbit/sec systems operating over 600 km using SMF or NZ-DSF. The system design for 2.5-Gbit/sec systems is exactly the same for SMF or NZ-DSF. The only difference in a 10-Gbit/sec design between SMF and NZ-DSF is the amount of DCF used for compensation. While NZ-DSF requires external dispersion compensation at the terminal locations, SMF also requires compensation at the intermediate points since it has three times the dispersion along the fiber. With 10-Gbit/sec systems, this additional DCF requirement is the key difference between SMF and NZ-DSF networks. A cost analysis is required to evaluate which configuration is more economical, since NZ-DSF fiber is priced at a significant premium to SMF. Currently, NZ-DSF is priced at more than two times the price of SMF. With this cost difference, SMF will be the preferred choice unless there are other system costs or technical considerations.

      For applications up to 2.5 Gbits/sec, system designs for SMF or NZ-DSF offer identical performance for DWDM applications. SMF is recommended because of significant fiber price savings.

      For applications at 10 Gbits/sec, DCF costs need to be considered. At less than 80 km, SMF is clearly recommended again since the system requirements for SMF and NZ-DSF are identical. For routes over 80 km, a present value economic analysis is required. SMF requires DCF compensation for distances greater than 80 km, and NZ-DSF requires compensation at distances exceeding 300 km. However, NZ-DSF requires less compensation per kilometer than SMF.

      Since DCF is only required when the fiber is upgraded or activated with a 10-Gbit/sec system, the DCF cost is likely deferred until it is activated with a 10-Gbit/sec system. Deferring the installation and cost of dispersion compensation has several benefits:

      • The cost of capital or interest reduces the current value cost of DCF deployed in the future.
      • DCF costs will likely drop in the future, enabling even greater savings.
      • Alternative dispersion-compensating technologies will likely become available that will further reduce the cost of compensation or enhance performance.

      Finally, technical limitations of NZ-DSF fiber at high data rates may prevent deployment of future DWDM equipment operating at higher power and narrow channel spacings, potentially requiring deployment of additional fiber cables as an alternative to achieve higher bandwidth targets. These future upgrade costs must be factored into any long-term net current value cost considerations.Th 07spr062

      Fig. 2. Shown are the savings provided with singlemode fiber when 10% of the fiber is deployed with 10-Gbit/sec systems. The remaining fibers are deployed with 10-Gbit/sec hardware at a rate of 10% every two years. Advances in dispersion compensation may improve these figures.

      With 10-Gbit/sec systems, SMF savings depend on the 10-Gbit/sec deployment rate. Figure 2 shows an example of SMF savings with 10% of the fiber initially deployed with 10 Gbits/sec. The remaining fibers are deployed with 10-Gbit/sec systems at a rate of 10% every two years. The savings range from approximately 10% to almost 50%, depending on the system length deployed. SMF also saves almost 50% over NZ-DSF when bit rates other than 10 Gbits/sec are used, regardless of system length.Th 07spr063

      Fig. 3. This figure shows the present value impact of DCF deployed in the future. The true cost (present value) of dispersion-compensating fiber drops significantly as time passes. Although not included in the figures displayed here, advances in dispersion compensation and future price decreses should make these costs decrease futher.

      An example of the expenditures and time value of money factors for a 240-km route length is shown in Figure 3. The true cost of DCF with SMF drops drastically for 10-Gbit/sec systems deployed in the future. These savings are even more significant if DCF price reductions or new technology costs are considered. For example, emerging innovations in fiber Bragg gratings or high-order mode conversion promise to improve external dispersion compensation in the next two to five years. As in Figure 2, this example shows savings with 10% of the fibers with 10-Gbit/sec systems at initial deployment and 10% of the remaining fibers are activated every two years assuming no reductions in DCF costs.

      The network installer should evaluate the total system cost over the project lifetime to compare current value costs of alternatives for a true understanding of the financial impact.

      With current system technologies, SMF and NZ-DSF have the same capabilities. SMF may require dispersion compensation with 10-Gbit/sec routes, depending on the route distance. This compensation investment is only required at the time the fiber is activated for 10-Gbit/sec transmission, thus enabling operators to defer spending.

      As there have been numerous technical drivers in the past that change the key system design-limiting factors, there will be new technology changes that will continue to affect fiber choice.

      The fiber choice for compatibility with future system drivers is dependent on the network technology outlook. If the future system growth is to be met with closer channel spacings or new amplifier bands (L or S), SMF best suppresses nonlinearity effects and will be capable of handling more channels. NZ-DSF applications with 50-GHz spacings are expected to be limited with 10-Gbit/sec systems. These limitations have been publicized in research findings by Alcatel and others at the recent Optical Fiber Communications Conference (OFC '99) in San Diego. In fact, several large network operators have recently selected SMF to futureproof their networks for emerging DWDM applications (see Fig. 4).Th 07spr064

      Fig. 4. Advances in both time-division multiplexing and DWDM technology promise a variety of choices for future high-speed networks. This chart illustrates the fiber types that should work best with each option.

      With a migration toward 40-Gbit/sec time-division multiplexing, both SMF and NZ-DSF will require external dispersion compensation. At these data rates, dispersion will become a critical limiting parameter for both types of fiber. DWDM fiber transmission limitations from nonlinearities will not be completely known until the optimal future transmission format is chosen. This will determine how closely channels will be spaced. Further studies are being conducted by various manufacturers and network operators to characterize future fiber needs at 40 Gbits/sec and beyond.

      Fiber selection for network installation requires an assessment of three key considerations. First, current system requirements must be evaluated. SMF and NZ-DSF are technically equivalent with current network transmission technologies. Both fiber types are capable of 700-km route spans with 2.5-Gbit/sec systems and up to 600 km with 10-Gbit/sec systems. Second, future upgrade requirements must be considered. SMF provides the best choice for future growth with narrower DWDM channel spacings due to suppression of nonlinearity effects. Current NZ-DSF products need to further optimize their dispersion and effective-area characteristics to compete with the close DWDM spacing performance of SMF. Finally, a project timeline of value cost analysis must be performed for 10-Gbit/sec applications to determine the cost tradeoffs over time.

      Therefore, SMF is currently the cost-effective solution for applications under 2.5 Gbits/sec. SMF is also a lower-cost option for 10-Gbit/sec applications up to 80 km. For distances past 80 km, the most economical fiber depends on the timing of 10-Gbits/sec deployment due to different DCF needs between fiber types.

      Jim Ryan is the new business development marketing manager for cable network products at Alcatel (Plano, TX). Questions on this article can be forwarded to Jim.Ryan@USA.Alcatel.com.

More in Network Design