Open Access
Add to BibSonomy
Abstract
Due to the cross phase modulation (XPM) effect, in long-haul high-speed dense wavelength division multiplexing (DWDM) coherent systems, using a low-speed on-off-keying (OOK) format optical supervisory channel (OSC) will introduce extra nonlinear phase noise, which restricts the transmission distance. In this paper, we propose a simple OSC coding method to mitigate the OSC-induced nonlinear phase noise. According to the split-step solution of the Manakov equation, we up-convert the baseband of the OSC signal out of the pass-band of the walk-off term to reduce the spectrum density of XPM phase noise. Experimental results show that the optical signal to noise ratio (OSNR) budget on the 400 G channel of 1280-km transmission is improved by 0.96 dB, which achieves almost the same performance with the no OSC case.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Optical supervisory channel (OSC) is essential in long-haul dense wavelength division multiplexing (DWDM) transmission systems for monitoring and controlling the transmission of each channel. The OSC usually uses on-off-keying (OOK) modulation at low bit rate ranging from 2 Mb/s to 1 Gb/s. However, it is reported that the OOK format channel imposes cross phase modulation (XPM)-induced nonlinear phase noise (NPN) on the coherent channels, which severely degrades the performance of the DWDM system [1–6].
To compensate the OSC-induced NPN, many digital signal processing (DSP) methods have been proposed. In [7], a longer low-density parity check (LDPC) code cooperating with innate deeper interleaving is reported to relieve the performance degradation caused by XPM. A novel polarization coupled carrier phase estimation (PCC) method for WDM coherent polarization-multiplexed systems is proposed in [8]. Since the XPM phase noise is correlated in the x- and y-polarization, the PCC algorithm employs polarization coupled information to mitigate the XPM-induced phase shifts. Although effective, the above-mentioned algorithms reveal relatively higher computational complexity. They manage to recover the original information from the received signal affected by XPM, rather than suppress the generation of XPM-induced NPN physically. To suppress the generation of NPN, one method is converting the OSC modulation format from OOK to binary phase shift keying (BPSK), which has much less XPM effect on the main channel. In [9], using an additional semiconductor optical amplifier (SOA), a continuous wave (CW) laser and optical filters at the optical cross-connect, the original OOK signal can be modulated on the phase of the CW and converted to BPSK signal. Nevertheless, the extra optical components will also increase the system cost and complexity, which is unsuitable for commercialization. In addition, the above researches focus on the 10 G class OOK and coherence mixed transmissions, whereas OSC usually uses low-cost and low-speed pluggable modules. Ref. [10] reported that the XPM-induced NPN increased with the decrease of OSC bit rate, which means in today's high-capacity transmission systems, an efficient and simple method to eliminate the XPM introduced by low-speed OSC is urgently needed, especially for the OSC with 100-200 Mb/s rate.
In this paper, we propose a simple coding method which can greatly reduce XPM-induced NPN by frequency shifting of low-speed OSC, according to the response characteristics of XPM. The proposed method is transparent to the DWDM channel format and rate. We also experimentally demonstrate a 400 G channel transmission loading with dual polarization probabilistic constellation shaping quadrature amplitude modulation (DP-PCS-16QAM) mixed with a few 100-Gb/s (DP-QPSK) channels over 1280-km standard single mode fiber (SSMF). Applying the proposed method will increase the optical signal to noise ratio (OSNR) budget of the system by 0.96 dB. BER performance of the proposed scheme remains the same as the case without OSCs.
2. Principle
The principle of the proposed coding method is based on the split-step solution of Manakov equation. According to the response characteristics of XPM, the OSC signal is up-converted to reduce XPM phase noise. In general, the vector optical propagation in communication fibers can be describe by Manakov equation [11–13]:
Fig. 1(a) shows an example of the frequency response of the walk-off term in one split-step of 2-km according to real link. The center frequency of two channels is set at 1511 nm and 1530 nm. The shape of walk-off term is similar to a low-pass filter. According to Eq. (2), it only transforms the lower frequency part of the power spectrum to the phase shift spectrum. The higher frequency part is automatically filtered and abandoned. Therefore, applying up-conversion to the OSC signal can move the main lobe out of the pass-band of the walk-off term. Since the main lobe is suppressed in the phase spectrum, the decrease of the spectrum density can help to reduce the XPM phase noise, as is presented in Fig. 1(b). The frequency of up-conversion should be designed according to the pass-band width of walk-off term response, which mainly depends on the dispersion of the fiber and the channel spacing. According to Eq. (2), moving the coherent channel away from the OSC, the pass-band of the walk-off term response narrows. Correspondingly, XPM phase noise is reduced. Therefore, when designing the up-conversion frequency, only the channel nearest to the OSC, which has the widest pass-band and most XPM phase noise, needs to be considered. If the OSC signal is moved out of the widest pass-band, the proposed scheme is also effective for channels further away.
Fig. 1. (a) Example of the walk-off term frequency response; (b) Schematic diagram of mitigating nonlinear penalty by digital up-conversion (DUC) scheme.
In the proposed scheme, compared with the traditional analog up-conversion, digital up-conversion is more flexible and simpler. It can be realized by XOR a high-speed 01 repeated sequence onto the original low-speed data sequence, as is shown in Fig. 2. The encoding and decoding process is similar to the direct sequence spread spectrum (DSSS). Figure 3 presents the spectrum of the OSC signals with original data sequence and encoded sequence. The up-conversion frequency can be set at as high as half the bit rate of the encoded sequence. After DUC, the DC component still remains in the OSC optical signal because of intensity modulation. Since the XPM phase shift generated by the DC part of the power signal is constant, the remaining optical carrier will not increase the XPM phase noise. Although the proposed scheme requires switching to higher-speed OSC optical modules, prices of 155M∼1 G 80-km commercial optical modules are merely about $\$$50-100. Therefore, the cost difference of OSC optical modules is negligible compared to the entire system.
In general, limited by the device bandwidth and fiber dispersion, applying a higher speed optical module instead will reduce the signal-to-noise ratio at the receiver end, which also reduces the transmission distance of the OSC channel per span. However, in the proposed DUC coding scheme, if the up-conversion frequency is $N$-times the bit rate of the original sequence, an original single symbol is represented by consecutive $2N$ symbols after encoding. Therefore, similar to the DSSS, the proposed coding method enhances the noise tolerance of the system. By adding white Gaussian noise to the original and encoded sequences, Fig. 4 shows the SNR gain of the coding methods. Twice and 4 times DUC coding have 3.5-dB and 6.5-dB SNR gain.
3. Simulation and discussion
In order to verify the feasibility of the proposed scheme, we simulate the XPM-induced phase shift in a span of 80-km SSMF using VPI Transmission Maker. Figure 5 shows the setup of the simulation system. It only contains a low-speed OOK OSC and a 90-Gbaud DP-16QAM coherent channel. The wavelengths of the OSC and the coherent channel are 1511 nm and 1530 nm, respectively. In this simulation, we focus on the nonlinear phase noise caused by XPM effect. To make XPM effect more significant, the launch power of the OSC is set at 8 dBm. Meanwhile, to separate the XPM-induced phase shift from the phase noise caused by the linewidth of the lasers, the linewidth and frequency offset of the coherent transceiver are both set to 0 Hz. Therefore, the frequency offset compensation and phase noise compensation in DSP of the receiver is abandoned. Considering the rate of the commercial optical modules used in the experiment and the times of up-conversion, the original rate of OSC is slightly different from the commonly used 155-Mb/s (OC-3 structure). The transmitted data of the OSC include 3 cases: original 125-Mb/s PRBS sequence, 250-MHz up-converted sequence (500 Mb/s) and 500-MHz up-converted sequence (1 Gb/s).
Constellations of the 16-QAM coherent channel after 80-km transmission are compared in different OSC cases. Figure 6 and Table 1 show the constellations and corresponding error vector magnitude (EVM). EVM describes the effective distance of the received complex symbol from its ideal position in the constellation diagram, which is suitable to reflect the introduced phase noise [15]. Compared with the case without OSC, adding an OOK OSC will introduce extra phase noise, which makes each point in the constellation spread to an “arc”, especially the outer ones. The OSC-induced phase noise also increases the EVM of the constellation from 2.98% to 16.95%. As is shown in Fig. 6(c) and (d), the proposed up-conversion coding scheme effectively suppresses the phase noise. The distribution of the signals in the constellations are still “points” rather than “arcs”. Compared with the case without OSC, the schemes with 250-MHz and 500-MHz up-converted OSCs have only 2.39% and 0.88% EVM degradation. The 500-MHz up-conversion scheme has the better performance, which is almost the same with no OSC case.
Fig. 6. Constellations of the 16-QAM coherent channel after 80-km transmission (a) without OSC, (b) with original OSC, (c) with 250-MHz up-converted OSC, (d) with 500-MHz up-converted OSC.
Table 1. EVM of the constellations of the four OSC cases
Figure 7 present the phase noise spectrum of the above four cases. Comparing the spectrums in Fig. 7(a) and (b), the extra phase noise caused by 125 Mb/s OSC is concentrated in the low frequency part. Since bandwidth of the baseband OSC signal is less than the passband width of the walk-off term, the power spectrum of the OOK signal is almost completely transferred to the phase noise spectrum, which is consistent with the inference from Eq. (2). After up-converting the baseband signal out of the passband width of the walk-off term, as is illustrated in Fig. 7(c) and (d), the main lobe of OSC signal is suppressed when transferred to phase spectrum. Most of the OOK signal energy is concentrated in the main lobe. Therefore, up-conversion method reduces the overall spectrum density of XPM phase noise.
Fig. 7. Phase noise spectrum of the four OSC cases (a) without OSC, (b) with original OSC, (c) with 250-MHz up-converted OSC, (d) with 500-MHz up-converted OSC.
In addition, higher frequency up-conversion leads to weaker remaining main-lobe and side-lobe. There is still quite a few main-lobe remaining in the 250-MHz up-conversion scheme. However, in the 500-MHz up-conversion scheme, the main-lobe is nearly suppressed completely. With less energy transferred from OOK power signal to the phase noise of the 16QAM signal, the 500-MHz scheme has better performance, which verify the feasibility and effectiveness of the proposed approach.
Higher rate OSC introduces less XPM phase noise. We also show the simulation results of the 16-QAM coherent channel with 1 Gb/s OSC using uncoded PRBS sequence. EVM of the constellation is 7.89%, which is much lower than that of OSC using 125-Mb/s PRBS sequence shown in Table 1. However, it is still inferior to the two DUC schemes. Comparing the phase noise spectrums in Fig. 8(b) and Fig. 7(d), since the energy of 1-Gb/s PRBS signal concentrates in the low frequency part, there exists more residual low-frequency phase noise than the DUC schemes. Therefore, the system performance improvement using DUC mainly comes from the designed OSC signal spectrum, not only from the higher OSC rate.
Fig. 8. (a) Constellation and (b) phase noise spectrum of the 16-QAM coherent channel with 1 Gb/s PRBS OSC.
In the practical coherent transmission systems, the Viterbi-Viterbi based phase noise compensation in DSP of the receiver end imposes a high-pass filter on the phase spectrum of the signal, which eliminates the low-frequency phase jitter caused by the laser linewidth. Meanwhile, it also filters out most of the low-frequency part of the XPM-induced phase shift. However, after long-haul transmission of many spans, the residual part of accumulated XPM-induced phase shift will still lead to severe performance degradation. It is necessary to be suppressed using the proposed up-conversion coding scheme. In Section 4, we experimentally demonstrate the effectiveness of the proposed scheme in long-haul fiber link.
4. Experiment and results
The experimental setup is illustrated in Fig. 9. We use commercial optical modules as the DWDM channels, and a 1-GBaud OOK optical modules as the OSCs (the original uncoded OSC rate is 125 Mb/s). The central wavelength of the OSCs is 1511.15 nm. The launch power of OSCs into the fiber is set to 5 dBm or 8 dBm. As is shown in Fig. 9(a), the optical transform unit (OTU) includes 94 channels. The first five of them are optical modules, and the rest are replaced by ASE noise. In the second channel at 1529.94 nm, we apply a 400 G DP-PCS-16QAM optical module as the measured channel. The baud rate of the module is set at 91.6 Gbaud. The others four 100 G optical modules use DP-QPSK modulation format, and the baud rate is 33.6 Gbaud. These four channels are used to simulate the interference of other channels in the actual link to the selected channel. The 1280-km fiber link consists of 16 spans of G.652.D fiber with a length of 80-km per span. Each span includes an EDFA to compensate the attenuation, and a fiber interface unit (FIU) to insert and remove OSC signals. Figure 9(b) shows the scene of field trial. In this experiment, we use field programmable gate array (FPGA) as the encoding circuit for OSC optical module, as shown in Fig. 9(c) and (d). Uncoded 125-Mb/s sequence is also generated by a 1-GBaud optical module via repeating each bit 8 times. The optical waveform and spectrum are almost the same with that generated by a low-speed module. Using FPGA and high-speed module to change the rate can help make sure that the modulation feature of modules won’t impact the results.
Fig. 9. (a) Experimental setup; (b) Scene of field trial; (c) OSC encoding circuit based on FPGA; (d) Original and encoded sequences of the FPGA outputs.
By adding OSCs by span, we measured the bit error ratio (BER) of the measured channel versus number of added OSCs with different configuration, as shown in Fig. 10. Using uncoded 125 Mb/s OSCs, the performance of the measured channel degrades rapidly with the increase of the number of OSCs. At 8-dBm launch power of OSC, the system can only support four added OSCs. After that, we digitally up-convert the OSC baseband to 250 MHz and 500 MHz. At the launch power of 5 dBm, the up-conversion schemes of 250 MHz and 500 MHz have 7.24 × 10−3 and 8.09 × 10−3 BER improvement respectively compared with uncoded OSC case at the spans of 16. Correspondingly, the reductions of EVM are 1.92% and 2.17%, as is shown in Table 2. Compared with no OSC case, the 500-MHz up-conversion scheme has only 0.31% EVM degradation. This indicates that it can essentially eliminate the XPM influence, achieving almost the same performance as no OSC case. At the launch power of 8 dBm, the proposed schemes can support adding all 16 OSCs. Compared with the 250-MHz up-conversion scheme, the 500-MHz scheme can better suppress the XPM phase noise. It has only 2.5 × 10−3 BER penalty compared with no OSC case. The corresponding increase of EVM is only 0.66%.
Table 2. EVM of the constellations at the receiver end
Table 3 shows the OSNR budget in each case. The launch power of the OSC is 5 dBm. OSNR budget is defined as the extra OSNR degradation that can be tolerated at the receiver end of the system when the FEC fails. Uncoded OSCs reduce OSNR budget by 1 dB compared with no OSC case. By contrast, the up-conversion schemes of 250 MHz and 500 MHz only induce 0.11 dB and 0.04 dB OSNR penalty to the uncoded OSC case, which are 0.89 dB and 0.96 dB lower than uncoded case.
Table 3. OSNR budget and penalty
In Table 4, we compare the OSNR penalties of different channels. Increasing the spacing of the OSC and coherent channel, the OSC-induced OSNR penalty is weakened. At 1560.2 nm, the OSCs hardly cause system performance degradation. The results show that the proposed DUC schemes can also mitigate the XPM phase noise of channels far away from 1529.9 nm. At 1544.1 nm, the up-conversion schemes of 250 MHz and 500 MHz reduce the OSNR penalty from 0.45 dB to only 0.13 dB and 0.05 dB, respectively.
Table 4. OSNR penalties of different channels
5. Conclusion
To reduce OSC-induced nonlinear phase noise in high-speed DWDM system, a simple up-conversion coding method is proposed for the low-speed OSCs. By digitally up-converting the baseband of the OSC signal out of the pass-band of the frequency response of walk-off term, the XPM phase noise is much suppressed. We experimentally demonstrate a 1280-km 400 G DP-PCS-16QAM transmission with OOK format OSCs. The original OSC bit rate is 125 Mbit/s. The results show that the OSNR budget can be improved by 0.96 dB using encoded OSCs, achieving almost the same performance as the case without OSCs.
Funding
National Key Research and Development Program of China (2021YFB1808200); National Natural Science Foundation of China (62275091).
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. H. Maeda, H. Kawahara, K. Saito, T. Seki, and J. Kani, “Performance Degradation of SD-FEC Due to XPM Phase Noise in WDM Transmission System with Low-Speed Optical Supervisory Channel,” in Proc. IEEE Photonics Conference (2019), pp. 1–2.
2. A. Bononi, M. Bertolini, P. Serena, and G. Bellotti, “Cross-Phase Modulation Induced by OOK Channels on Higher-Rate DQPSK and Coherent QPSK Channels,” J. Lightwave Technol. 27(18), 3974–3983 (2009). [CrossRef]
3. O. Vassilieva, T. Hoshida, J. C. Rasmussen, and T. Naito, “Symbol Rate Dependency of XPM-induced Phase Noise Penalty on QPSK-based Modulation Formats,” in Proc. Eur. Conf. Opt. Commun (2018), paper We1E.4.
4. B. Spinnler, N. Hecker-Denschlag, S. Calabro, M. Here, C.-J. Weiske, E.-D. Schmidt, D. van den Borne, G.-D. Khoe, H. de Waardt, R. Griffin, and S. Wadsworth, “Nonlinear tolerance of differential phase shift keying modulated signals reduced by XPM,” in Proc. Optical Fiber Communication Conference (2004), paper TuF3.
5. M. S. Alfiad, M. Kuschnerov, T. Wuth, et al., “111-Gb/s Transmission Over 1040-km Field-Deployed Fiber With 10 G/40 G Neighbors,” IEEE Photon. Technol. Lett. 21(10), 615–617 (2009). [CrossRef]
6. W. Astar, A. S. Lenihan, and G. M. Carter, “Performance of DBPSK in a 5 /spl times/ 10 Gb/s mixed modulation format Raman/EDFA WDM system,” IEEE Photon. Technol. Lett. 17(12), 2766–2768 (2005). [CrossRef]
7. P. Leoni, V. Sleiffer, S. Calabrò, V. Veljanovski, M. Kuschnerov, S. Jansen, and B. Lankl, “Impact of Interleaving on SD-FEC Operating in Highly Non-Linear XPM-Limited Regime,” in Proc. Optical Fiber Communication Conference/National Fiber Optic Engineers Conference (2013), paper OW1E.6. [CrossRef]
8. K. Piyawanno, M. Kuschnerov, F. N. Hauske, M. S. Alfiad, B. Spinnler, A. Napoli, H. de Waardt, and B. Lankl, “Polarization coupled carrier phase estimation for coherent polarization multiplexed QPSK with OOK-neighbours,” in Proc. Optical Fiber Communication Conference (2009), paper OMT6. [CrossRef]
9. W. Astar and G. M. Carter, “Mitigation of XPM Penalty in 10-Gb/s OOK/DBPSK Mixed Modulation Format 50-GHz-Spaced WDM Transmission by Conversion of OOK to BPSK Using an SOA,” IEEE Photon. Technol. Lett. 20(20), 1715–1717 (2008). [CrossRef]
10. H. Kawahara, K. Saito, T. Seki, T. Kawasaki, and H. Maeda, “ The Impact of Nonlinear Phase Noise Induced from Low-Speed Optical Supervisory Channel on Soft-Decision FEC Performance,” in Proc. Optical Fiber Communication Conference (2020), paper Th2A.53.
11. F. P. Guiomar, F. P. Guiomar, S. B. Amado, R. M. Ferreira, J. D. Reis, S. M. Rossi, A. Chiuchiarelli, J. R. F. de Oliveira, A. L. Teixeira, and A. N. Pinto, “Multicarrier Digital Backpropagation for 400 G Optical Super-channels,” J. Lightwave Technol. 34(8), 1896–1907 (2016). [CrossRef]
12. J. Leibrich and W. Rosenkranz, “Efficient numerical simulation of multichannel WDM transmission systems limited by XPM,” IEEE Photon. Technol. Lett. 15(3), 395–397 (2003). [CrossRef]
13. E. F. Mateo, F. Yaman, and G. Li, “Efficient compensation of inter-channel nonlinear effects via digital backward propagation in WDM optical transmission,” Opt. Express 18(14), 15144–15154 (2010). [CrossRef]
14. E. F. Mateo, X. Zhou, and G. Li, “Improved digital backward propagation for the compensation of inter-channel nonlinear effects in polarization-multiplexed WDM systems,” Opt. Express 19(2), 570–583 (2011). [CrossRef]
15. R. Schmogrow, B. Nebendahl, M. Winter, A. Josten, D. Hillerkuss, S. Koenig, J. Meyer, M. Dreschmann, M. Huebner, C. Koos, J. Becker, W. Freude, and J. Leuthold, “Error Vector Magnitude as a Performance Measure for Advanced Modulation Formats,” IEEE Photon. Technol. Lett. 24(1), 61–63 (2012). [CrossRef]
