RSoft Enewsletter

September 2018

Terrestrial Optical Fiber Networks for Long-Haul Data Center Connectivity: Engineering Optimal Spans

Span engineering in legacy, intensity-modulated, long-haul wavelength division multiplexed (WDM) networks typically involves optimal choice and placement of dispersion compensating units (DCUs) and line amplification schemes to overcome linear impairments of the optical transport. The rapid emergence of big data, the Internet of Things (IoT), cloud services, video, eCommerce, and social media have resulted in a dramatic explosion of data and Internet traffic in recent years, forcing a move to spectrally efficient, advanced modulation formats with coherent reception. Consequently, dispersion compensation is delegated to the receiver, while improved detection sensitivity and spectral efficiency can support transmission up to a few thousand kilometers with minimal amplification. As a result, the focus of span engineering has moved from the management of linear impairments to the management of nonlinearities due to the optical transport. While OptSim can help optimize spans for both scenarios, this article focuses on the latter.

Figure 1 shows an OptSim schematic of a 9000-km long, 9-channel WDM system with a PM-QPSK data rate of 137 Gbps (34.25GBd/s) per channel, including forward error correction (FEC) overheads. The dynamic receiver uses a training sequence and dynamic tracking for equalization. OptSim’s least mean squares (LMS) algorithm is used to calculate the initial equalization coefficients during the training, as well as to update their values during the decision-driven phase. After the constant modulus algorithm (CMA) is applied to estimate the channel’s Jones matrix, the phase is recovered using the Viterbi-Viterbi algorithm. Bit error rate (BER) is obtained by direct counting of errors.

Figure 1. OptSim schematic of a 9000-km long, 9-channel coherent fiber-optic transmission

Figure 2 shows back-to-back waveforms at the transmitter and at the receiver for the center channel. 

Figure 2. Back-to-back waveforms at the transmitter and receiver for the center channel

As a side note, this type of simulation is inherently time consuming for a number of reasons: (i) higher data rates and channel-count; (ii) thousands of kilometers of transmission distances; (iii) a large number of transmitted bits to facilitate BER counting; (iv) accumulated dispersion and nonlinearities; and (v) offline DSP. You can use OptSim’s Gaussian Noise (GN) emulator model to replace the fiber, which can reduce the simulation time by one third. It is shown experimentally [1], via OptSim-based modeling [2], and analytically [3] that for uncompensated, long-haul coherent systems, nonlinear transmission impairments in fiber manifest as nonlinear interference (NLI) with Gaussian noise power spectral density.

The simulation was carried out to obtain BER versus per channel transmitted power curves (bell curves) for two span configurations: (i) 60 spans with 150-km fiber per span; and (ii) 150 spans with 60-km fiber per span. Figure 3 shows a performance comparison for both span configurations.

Figure 3. Comparison of two different configurations of spans

As we can see, for the same total transmission distance of 9000-km, configuration (i) with higher accumulated nonlinearities (due to longer fiber per span) gives optimal per-channel power of around 3dBm, while configuration (ii) gives optimal per-channel power of around -2dBm with a much lower BER.

Since optically uncompensated spans in coherent systems are shown to perform better than dispersion-managed transmission [4], green-field deployments are better off having spans with optically uncompensated transmission. Since dispersion helps reduce NLI-induced penalty, you may find it an interesting exercise to study green-field installation via simulations in OptSim with different types of singlemode fibers (SMF) in order to make an optimal choice. For example, Ref. [2] compares pure-silica-core fiber (PSCF), standard SMF, and large effective-area nonzero-dispersion shifted fiber (NZDSF), while Ref. [5] compares seven different fiber types, and Ref. [6] provides design rule analysis for maximum reach based on experimental studies for different fiber types.

For more information or to discuss your photonic integration modeling needs, please contact rsoft_support@synopsys.com  

References:

  1. Torrengo, E., et al., “Experimental validation of an analytical model for nonlinear propagation in uncompensated optical links,” Optics Express, vol. 19, no. 26, 2011.
  2. Carena, A., et al., “Modeling of the impact of nonlinear propagation effects in uncompensated optical coherent transmission links,” Journal of Lightwave Technology, vol. 30, no. 10, 2012, pp. 1524-1539.
  3. Poggiolini, P., et al., “The GN-model of fiber non-linear propagation and its applications,” Invited Paper, Journal of Lightwave Technology, vol. 32, no. 4, 2014, pp. 694-721
  4. V. Curri, P. Poggiolini, A. Carena, and F. Forghieri, “Dispersion compensation and mitigation of non-linear effects in 111 Gb/s WDM coherent PM-QPSK systems,” IEEE Photon. Technol. Lett., vol. 20, no. 7, pp. 1473–1475, Sep. 1, 2008.
  5. A. Nespola, et al., “GN-model validation over seven fiber types in uncompensated PM-16QAM Nyquist-WDM links,” IEEE Photon. Technol. Lett., vol. 25, no. 2, pp.206 -209, Jan. 5, 2014.
  6. V. Curri, et al., “Design rules for reach maximization in uncompensated Nyquist-WDM links,” Proc. of 39th European Conference on Optical Communication (ECOC) 2013, pp. 717-719, September 2013.

Upcoming Fall 2018 Training

We are taking registrations for our upcoming RSoft products training sessions. Visit our website for more information, and to register online today. Seats are limited, minimum enrollments apply.

RSoft Component Tools Training

October 22-24, 2018
Synopsys, Inc.
400 Executive Boulevard, Suite 101
Ossining, NY 10562

RSoft System Tools Training

October 25-26, 2018
Synopsys, Inc.
400 Executive Boulevard, Suite 101
Ossining, NY 10562

If you would like customized, on-site training at your company, please email rsoft_support@synopsys.com. We can tailor the class to your specific needs.