The transmission performance of coherent dual-polarization multi-band OFDM (DP-MB-OFDM) and QPSK (DP-QPSK) are experimentally compared for 100 Gb/s long-haul transport over legacy infrastructure combining G.652 fiber and 10 Gb/s WDM system. It is shown that DP-MB-OFDM and DP-QPSK have nearly the same performance at 100 Gb/s after transmission over a 10 × 100-km fiber line. Furthermore, the origin of performance degradations and limitations of the DP-MB-OFDM is explored numerically, as well as the impact of transmission distance and sub-band spacing.
©2013 Optical Society of America
The popularity of bandwidth-consuming applications such as internet video, cloud storage and social networking requires large volumes of data to be transmitted over long distances. This has fueled the exponential growth of data traffic volumes in the telecommunication networks, and based on this current trend it will likely continue to drive an unsurpassed need for transmission capacity over the next decade. In this context, 100 Gb/s WDM transmission technologies are key to provide more capacity in optical long-haul transport networks.
Orthogonal frequency-division multiplexing (OFDM) is an ultra-high spectral efficiency technology able to eliminate inter-symbol interference (ISI) caused by chromatic dispersion (CD) and polarization-mode-dispersion (PMD) . By using coherent OFDM (CO-OFDM) and multi-band strategy, Jansen et al.  demonstrated 121.9 Gb/s multi-band OFDM (MB-OFDM) WDM transmission over 1000 km of standard single-mode fiber (SSMF) with 2 b/s/Hz spectral efficiency. Furthermore, Coherent MB-OFDM [3, 4] is considered as a potential and credible candidate for wavelength-division multiplexing (WDM) transmission at 400 Gb/s and 1 Tb/s. On the other hand, coherent dual-polarization quaternary phase-shift keying (DP-QPSK) is today the industrial solution for 100 Gb/s long-haul transport . In particular, it has the valuable advantage to permit the smooth upgrade at 100 Gb/s of the legacy 10 Gb/s long-haul transport infrastructures [6, 7] thanks to its compatibility with 50 GHz channel spacing and its high robustness to PMD.
However, a debate exists over the capacity of OFDM to be as efficient as QPSK for long-haul WDM transmission due to its supposed higher sensitivity to fiber nonlinearities. After having answered to this question in  for the particular case of dispersion-compensation-fiber-free (DCF-free) transmission line, the case of legacy transport infrastructure mixing G.652 fiber and 10 Gb/s WDM system is addressed in this paper. An experimental performance comparison of DP-MB-OFDM and DP-QPSK at 100 Gb/s is carried out under realistic transmission conditions. Through the simultaneous propagation of two DP-MB-OFDM and DP-QPSK channels combined with 78 × 10 Gb/s nonreturn-to-zero (NRZ) on-off keying (OOK) channels in a 50 GHz-spaced WDM system using a 10 × 100-km dispersion-managed (DM) G.652 fiber line, it is shown that DP-MB-OFDM format is nearly as robust as DP-QPSK for 100 Gb/s transmission. In addition, the origin of performance degradations and limitations of the DP-MB-OFDM is explored numerically, as well as the impact over the system performance of transmission distance and sub-band spacing.
2. System model and transceiver set-up
The nominal data rate of 100 Gb/s, increased up to 124.4 Gb/s to account for the various transmission overheads (7% for forward error correction [FEC], 7.03% for cyclic prefix [CP], 6% for training symbols, and 2.3% for pilot tones), is split between four polarization-multiplexed OFDM sub-bands. Each OFDM sub-band can support 31.1 Gb/s in a bandwidth of ~8 GHz while the sub-band spacing is set at 10 GHz. Figure 1 illustrates the block diagram of a single-band dual-polarization OFDM digital transceiver. For simplicity, a perfectly synchronized system is assumed in this illustration, meaning that the carrier frequency offset (CFO)  and local oscillator (LO) phase noise  is compensated. In Fig. 1, the pseudo-binary random sequence (PRBS) is parallelized and mapped into QPSK symbols. Six pilot tones are dedicated to phase noise compensation and training symbols are inserted. A first sequence of training symbols [0 TS1 0] is added at the frame beginning for timing synchronization and CFO estimation. A second sequence of five training symbols with the following structure [0 TS2 0 TS2 0] is also periodically inserted in order to carry out polarization separation. Subsequently, the signal is modulated onto orthogonal subcarriers by employing the inverse fast Fourier transform (IFFT) of size 256. The highest 80 frequencies of the 256 subcarriers are set to zero in order to separate the OFDM base-band signal from the aliasing products generated at the output of the digital-to-analogue converters (DACs). These last ones are used at a sampling frequency of 12 GSa/s. Afterwards, the CP of 18 samples (1.4 ns) is added at the beginning of each OFDM symbol to mitigate impairments caused by, for instance, CD or PMD. The OFDM signals at the DAC outputs fed the optical front end. At the receiver, the outputs of the optical front end are first digitalized using analogue-to-digital converters (ADCs). The “off-line” digital signal processing (DSP) is then performed with four basic steps: synchronization according to the algorithm developed by Minn & Bhargava , which also permits to compensate a frequency offset in the range of ± 2∆f (∆f being the sub-carrier spacing); compensation of the remaining part of the frequency offset by determining the frequency shift that the last filled OFDM sub-carrier experiences ; separation of the two polarization components thanks to the zero-forcing (ZF) MIMO equalizer ; and finally compensation of the common phase noise generated by the external cavity laser (ECLs) thanks to the method of the pilot subcarriers described in . In the DP-QPSK case, the “off-line” DSP is based on blind adaptive equalization, and more particularly on theconstant modulus algorithm (CMA) which carries out polarization separation and residual CD compensation . Frequency offset compensation and carrier phase estimation is done by the methods described in . To avoid cycle slips, the detection is differential. Finally, these symbols are demodulated and the bit-error rate (BER) is calculated.
2.1 Experimental set-up
Figure 2 shows the experimental set-up of the employed 100 Gb/s DP-MB-OFDM transmitter. A comb of optical carriers spaced by 10 GHz (shown in the first inset of Fig. 2) is generated by using an ECL and driving a dual-arm Mach-Zehnder modulator (MZM) with a 10 GHz RF frequency according to the recommendations of . The required four optical carriers are selected at the transmitter output by a square flat-top optical band-pass filter (BPF) of ~40 GHz bandwidth. Before that, a combination of 20 GHz and 40 GHz polarization-maintaining delay line interferometers (PM-DLI) splits into four groups of carriers spaced by 40 GHz the initial comb of 10 GHz-spaced optical carriers. Each of the four generated combs is modulated by a complex-MZM (CMZM) and combined by a 4:1 polarization-maintaining (PM) coupler. The details over the generation of the OFDM signal are given in . Thanks to two arbitrary waveform generators (AWG), data carried by neighboring sub-bands are totally de-correlated, provided that AWG 1 generates the first and third sub-bands while AWG 2 generates the second and fourth sub-bands. A polarization-maintaining Erbium-doped fiber amplifier (PM-EDFA) balances the losses introduced by the MZM, DLIs, CMZMs and coupler and feeds a 1-symbol-delay polarization-multiplexing module. The spectrum of our 124 Gb/s DP-MB-OFDM operating at 1552.93 nm is shown in the second inset of Fig. 2 which is combined with one channel at 1548.11 nm carrying a 112 Gb/s DP-QPSK signal and a multiplex of 78 wavelengths spaced by 50 GHz and modulated at 10.7 Gb/s by NRZ-OOK format. The DP-QPSK channel is fed by an ECL to limit the impact of laser phase noise, while the 78 NRZ channels at 10.7 Gb/s are fed by standard laser diodes (LD) with wavelengths ranging from 1529.16 nm to 1560.61 nm. De-correlated 215-1 PRBSs at 28 Gb/s are used to drive the I and Q ports of the CMZMs, which generate QPSK constellation. A 10-ns timing delay is introduced between the two replica of the QPSK signal into the polarization-multiplexing module in order to generate the DP-QPSK signal. The 78 odd and even 10 Gb/s channels are firstly encoded thanks to a Reed Solomon (RS)-[255, 239] FEC code, separately multiplexed, independently modulated with de-correlated 231-1 PRBS, and then coupled with the two 100 Gb/s channels.
The dispersion-managed (DM) transmission line shown in Fig. 3 is constituted of a pre-compensation stage of −1000 ps/nm at 1550 nm, followed by ten spans of 100 km of G.652 SSMF, separated by double-stage EDFAs with 30 dB gain and 5.5 dB noise figure (NF), whose inter-stage is equipped with dispersion compensation module (DCM) adapted to 90 km SSMF spans. The optical power injected into the DCM was fixed to −10 dBm per channel. In the middle of our transmission line, a dynamic gain equalizer (DGE) is inserted in order to flatten the multiplex power after 500 km. A post-compensation stage of −700 ps/nm brings back to ~0 ps/nm the cumulated dispersion of the channel at 1550 nm. At the receiver side, the 100 Gb/s DP-MB-OFDM and DP-QPSK signals are selected by a square flat-top optical band-pass filter (OBPF) of 0.4 nm bandwidth, and detected by a polarization diversity coherent receiver using a ~100 kHz bandwidth ECL as LO. The signals are converted back to the digital domain thanks to a 50 GSa/s real-time digital phosphor oscilloscope (DPO). In the DP-MB-OFDM detection case, the LO wavelength is tuned to the centre of the OFDM sub-band under measurement. The 10.7 Gb/s NRZ channel under measurement is selected by a Gaussian OBPF of 0.25 nm bandwidth, detected by a 10 GHz photo-receiver, which feds the FEC decoder and 10.7 GHz clock and data recovery (CDR). The FEC card sends the decoded 231-1 PBRS to a 9.95 Gb/s BER tester (BERT).
2.2 Simulation set-up
For the theoretical investigations, the DP-MB-OFDM transceiver as well as the optical transmission is implemented using a Matlab/VPI-transmission-Maker co-simulated environment. It should be noted that, ideal coherent DP-MB-OFDM transceiver is considered here, except the phase-noise which has been included in simulations. The DAC/ADC clipping ratio and quantization are taken into account and set to 13 dB and 10-bits, respectively, which have no impact on the performance of OFDM signals for subcarrier number higher than 32 . Similar observations occur when no DAC/ADC clipping ratio and quantization bits are considered . The generated PRBS counts 338,000 bits sent over 1000 symbols by using 169 subcarriers per sub-band. Ten noise trials for each band are considered hereafter and an average BER is calculated for the four bands. In addition, the transmission distance varies from 800 km up to 1200 km, while the adopted modulation format is QPSK. In Table 1, the SSMF/DCF characteristics and transceiver parameters for the theoretical model are depicted.
3. Transmission performance and origin of performance degradations/limitations of DP-MB-OFDM
Two various configurations have been evaluated in Fig. 4. In the first one (“Single-Channel” configuration) used as reference, the wavelengths were not modulated except the channels at 1552.93 nm and 1548.11 nm, respectively, which carry the 100 Gb/s channels. The second one (“with 10G” configuration) corresponds to the experimental set-up already described previously, for which several schemes have been investigated. Firstly, no guard band (GB) is inserted between our two 100 Gb/s channels and the 10 Gb/s multiplex (“No GB” configuration). As observed in Fig. 4 for this scheme, the 100 Gb/s DP-QPSK and DP-MB-OFDM channels do not operate after 10 × 100-km of G.652 fiber, as BERs largely exceed the FEC limit (fixed here at 2 × 10-3).
In order to limit the impact of cross-phase modulation (XPM) and its cross-polarization modulation (XPolM) corollary between the 10 Gb/s NRZ channels and the 100 Gb/s ones , a first scheme consisting in inserting a guard band of 100 GHz and 150 GHz, corresponding to one and two 10 Gb/s channels stopped from each side of the measured 100 Gb/s channels (“GB=100 GHz” & “GB=150 GHz” configurations), have been tested. This option slightly improves the BER of 100 Gb/s channels while increasing the optimum span input power per channel of both 100 Gb/s DP-QPSK (Fig. 4(a)) and DP-MB-OFDM (Fig. 4(b)), but not sufficiently to be below the FEC limit. Note as well that increasing the GB width from 100 GHz to 150 GHz does not further improve transmission performance.
Then, provided that the 10 Gb/s NRZ channels have an advantage of 5 dB in terms of back-to-back OSNR sensitivity over both 100 Gb/s DP-MB-OFDM and DP-QPSK signals as shown in Fig. 5, we have intentionally reduced the 10 Gb/s channel power of 5 dB with respect to that of the 100 Gb/s signals (“P(100G)-P(10G) = 5 dB” configuration) into the transmitter. This solution improves significantly the BER of 100 Gb/s DP-QPSK and DP-MB-OFDM: nearly one BER decade gain is observed. In the same time, the optimum span input power per channel is enhanced of 3 dB for DP-QPSK and 2 dB for DP-MB-OFDM when compared to the “No GB” configuration. This configuration permits to recover system margins and to ensure an error-free 1000 km transmission, confirming that the 10 Gb/s channel power reduction option is the most credible solution to limit the impact of XPM and XPolM in such legacy infrastructure. Nonetheless, the BER obtained in the “Single-Channel” configuration is not recovered, in which intra-channel nonlinearities only were excited. Note that no error has been detected for the two 10 Gb/s nearest neighbours of the 100 Gb/s DP-MB-OFDM and DP-QPSK channels (at the optimum span input power per channel over the curves of Fig. 4). Figure 4 also points out that 100 Gb/s DP-QPSK shows a slightly higher performance than 100 Gb/s DP-MB-OFDM, which is less than half a BER decade in the various configurations under study. The optimum span input power per channel is also slightly higher (up to 1 dB) in the DP-QPSK case, indicating an upper resistance of DP-QPSK over DP-MB-OFDM to XPM and XPolM.
4. Numerical simulations predictions for the performance of 100 Gb/s DP-MB-OFDM
4.1 Simulation results for the 10x100-km dispersion-managed G.652 fiber line
Figure 6 summarizes the results: the BER as a function of the span input power per channel after 1000 km of transmission is plotted to investigate the origin of DP-MB-OFDM signal degradation by including / excluding inter-band nonlinear effects (i.e. only a unique OFDM sub-band is generated in the case “without inter-band nonlinear effects”). We can firstly observe that our experimental measurements match well with numerical predictions, confirming the validity of the developed numerical model. Secondly, a BER difference of two decades as well as an increase of the optimal span input power of ~1 dB is observed between the two cases under investigation, demonstrating a significant sensitivity of the multi-band OFDM configuration to cross-nonlinearities. The latter observation is directly related to the exacerbation of four-wave mixing (FWM) arisen from the inter-band crosstalk .
4.2 Impact of transmission distance and inter-band nonlinear crosstalk over the 100 Gb/s DP-MB-OFDM system performance
The origin of performance degradation of the adopted 100 Gb/s DP-MB-OFDM system is explored thoroughly under different noise conditions. In Fig. 7, the numerically calculated spectra of the DP-MB-OFDM signal after 1000 km is plotted for PIN SPAN ~0 dBm and the two following cases: (a) when only amplified spontaneous emission (ASE) noise is taken into account, and (b) when only Kerr effects are considered. The received constellation diagrams (2nd sub-band for X-polarization) corresponding to the two aforementioned cases with (a) BERa~5.9×10-6 and (b) BERb~2.96×10-6 are depicted in Fig. 7. A comparison between Fig. 7(a) and Fig. 7(b) shows that ASE noise and Kerr effects contribute identically to the BER degradation. The out-of-band distortions of the DP-MB-OFDM spectrum in Fig. 7(b) correspond to the FWM-induced new frequencies, pointing out the importance of FWM crosstalk to signal performance degradation.
The transmission performance of our 100 Gb/s DP-MB-OFDM system is explored numerically for various transmission distances: from 800 km up to 1200 km. In Fig. 8(a), the BER as a function of the total span input power per channel for such transmission distances is depicted. It is shown firstly, that, increasing the transmission distance results in a larger accumulation of ASE noise and fiber nonlinearities, degrading subsequently the BER. Secondly, it is revealed that transmission of 100 Gb/s DP-MB-OFDM over 1200 km (“triangles” curve) is error-free as the BER at the optimum span input power per channel (i.e. ~-2 dBm) is just below the FEC limit (fixed here at 2 × 10−3).
One simple method to reduce the inter-band nonlinear crosstalk of the DP-MB-OFDM system consists in increasing the sub-band spacing. In Fig. 8(b), the impact of sub-band spacing in the 100 Gb/s DP-MB-OFDM transmission system is numerically investigated over 1000 km. The sub-band spacing of the DP-MB-OFDM is increased from 10 GHz up to 25 GHz. As expected, increasing the sub-band spacing results to a BER improvement, in particular in the nonlinear regime (in the right-hand part of the BER vs. PIN SPAN curves). It should be noted that the BER improvement corresponding to values of PIN SPAN < −1 dBm (ASE noise dominant regime) is due to the non-neglectable inter-band nonlinear effects. This phenomenon is also observed in Fig. 6, in which the BER difference is revealed between the cases of including and excluding the inter-band Kerr-effect. In Fig. 8(b), the BER improvement in the nonlinear region induced by the broadening of the sub-band spacing from 10 GHz up to 25 GHz is around 33%: it is obvious that the effect of FWM crosstalk across the largely spaced sub-bands is then significantly reduced.
In this paper, we have shown that 100 Gb/s DP-QPSK and DP-MB-OFDM transmission over legacy fiber and system infrastructure is error-free after 10 × 100-km of G.652 fiber, under the condition to decrease by 5 dB the 10 Gb/s channel power with respect to the 100 Gb/s channel ones. In such realistic conditions, 100 Gb/s DP-MB-OFDM and DP-QPSK have nearly the same performance. From numerical investigations, we can conclude that an error-free transmission over 1200 km is feasible for the 100 Gb/s DP-MB-OFDM system. Moreover, it has been verified that DP-MB-OFDM is sensitive to the nonlinear inter-band crosstalk effects induced by FWM. A first method for reducing the inter-band crosstalk effects of the DP-MB-OFDM and to improve the system performance is to increase the sub-band spacing, but at the expense of the system spectral efficiency. Finally, it should be noted that the accordance between our numerical investigations and experimental measurements confirms the validity of the numerical model developed here.
This work has been supported by the “100G-FLEX” project of the “Pôle de Compétitivité Images & Réseaux” and by the “100GRIA” project of the “Pôle de Compétitivité Systematic”.
References and links
1. S. L. Jansen, I. Morita, T. C. W. Schenk, N. Takeda, and H. Tanaka, “Coherent Optical 25.8-Gb/s OFDM Transmission Over 4160-km SSMF,” IEEE J. Lightw. Technol. 26(1), 6–15 (2008). [CrossRef]
2. S. L. Jansen, I. Morita, T. C. Schenk, and H. Tanaka, “121.9-Gb/s PDM-OFDM Transmission With 2-b/s/Hz Spectral Efficiency Over 1000 km of SSMF,” IEEE J. Lightw. Techn. 27(3), 177–188 (2009). [CrossRef]
3. H. Takahashi, S. L. Jansen, A. A. Amin, I. Morita, and H. Tanaka, “Comparison between Single-band and Multi-band Optical OFDM at 120-Gb/s,” in Proc. Internat. Conf. on Opt. Internet (COIN 2008).
4. W. Shieh, “OFDM for flexible high-speed optical networks,” IEEE J. Lightw. Techn. 29(10), 1560–1577 (2011). [CrossRef]
6. J. Renaudier, O. Bertran-Pardo, H. Mardoyan, P. Tran, G. Charlet, S. Bigo, M. Le François, B. Lavigne, J.-L. Augé, L. Piriou, and O. Courtois, “Performance comparison of 40G and 100G coherent PDM-QPSK for upgrading dispersion managed legacy systems,” in Proceedings of the Optical Fiber Communication Conference (IEEE, 2009), Paper NWD5.
7. E. Pincemin, J. Karaki, M. Selmi, D. Grot, T. Guillossou, C. Gosset, Y. Jaouën, and P. Ciblat, “100 Gbps DP-QPSK performance over DCF-free and legacy system infrastructure,” in Proceedings of the IEEE Photonics Conference (2012), Paper TuE2.
8. J. Karaki, E. Pincemin, D. Grot, T. Guillossou, Y. Jaouën, and R. Le Bidan, “Multi-Band OFDM versus Single-Carrier DP-QPSK for 100 Gbps Long-Haul WDM Transmission,” in Proc. Signal Processing in Photonic Communications (SPPCom), Colorado Springs, CA, (2012), Paper SpTu1A.2.
9. T. M. Schmidl and D. C. Cox, “Robust frequency and timing synchronization for OFDM,” IEEE Trans. Commun. 45(12), 1613–1621 (1997). [CrossRef]
10. T. C. W. Schenk, RF Imperfections in High-Rate Wireless Systems (Springer, 2008).
11. H. Minn, M. Zeng, and V. K. Bhargava, “On timing offset estimation for OFDM systems,” IEEE Commun. Lett. 4(7), 242–244 (2000). [CrossRef]
12. W. Shieh, X. Yi, and Y. Tang, “Transmission experiment of multi-gigabit coherent optical OFDM systems over 1000 km SSMF fibre,” IEEE Electron. Lett. 43(3), 183–184 (2007). [CrossRef]
13. S. L. Jansen, I. Morita, T. C. W. Schenk, and H. Tanaka, “Long-haul transmission of 16×52.5 Gb/s polarization-division multiplexed OFDM enabled by MIMO processing,” OSA Journal of Optical Networking 7(2), 173–181 (2008). [CrossRef]
14. X. Yi, W. Shieh, and Y. Tang, “Phase estimation for coherent optical OFDM,” IEEE Photon. Technol. Lett. 19(12), 919–921 (2007). [CrossRef]
16. D.-S. Ly-Gagnon, S. Tsukamoto, K. Katoh, and K. Kikuchi, “Coherent detection of optical quadrature phase-shift keying signals with Carrier phase estimation,” IEEE J. Lightw. Techn. 24(1), 12–21 (2006). [CrossRef]
17. T. Sakamoto, T. Kawanishi, and M. Izutsu, “Asymptotic formalism for ultraflat optical frequency comb generation using a Mach-Zehnder modulator,” Opt. Lett. 32(11), 1515–1517 (2007). [CrossRef] [PubMed]
18. J. Karaki, E. Pincemin, Y. Jaouën, and R. Le Bidan, “Frequency offset estimation in a polarization-multiplexed coherent OFDM system stressed by chromatic dispersion and PMD,” in Proceedings of the Conference of Lasers and Electro-Optics (CLEO), (OSA, 2012), Paper CF1F.3.
19. E. Giacoumidis, J. L. Wei, X. L. Yang, A. Tsokanos, and J. M. Tang, “Adaptive modulation-enabled WDM impairment reduction in multi-channel optical OFDM transmission systems for next generation PONs,” IEEE Photon. Journal 2(2), 130–140 (2010). [CrossRef]
20. E. Giacoumidis, J. Karaki, E. Pincemin, C. Gosset, R. Le Bidan, E. Awwad, and Y. Jaouën, “100 Gb/s coherent optical polarization multiplexed Multi-band-OFDM (MB-OFDM) transmission for long-haul applications,” in Proceedings of the International Conference on Transparent Optical Networks (ICTON), (IEEE, 2012), Paper We.B1.2.