We propose 4D modulation with direct detection employing a novel Stokes vector Kramers Kronig transceiver. It shows that employing a Stokes vector receiver, transmitting a digital carrier and using Kramers Kronig detection offer an effective way to de-rotate polarization multiplexed complex double side band signal without using a local oscillator at receiver. The impact of system parameters and configurations including carrier-to-signal-power ratio, guard band of the digital carrier, digital oversampling ratio and real MIMO is experimentally investigated. Finally, a record net data rate of 400 Gb/s (a raw rate of 480 Gb/s including FEC overhead) over 80 km SSMF is achieved in a 60 Gbaud PDM-16QAM single carrier experiment with a BER below the threshold of 2.0x10−2.
© 2017 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Fiber optic communication traditionally supports the evolution of the Internet through vast backbone networks spanned over hundreds and thousands of kilometers around the globe. Today, long-haul optical transmission is mainly based on high speed coherent technology, which can achieve 400 Gb/s and beyond per wavelength [1,2]. In the past five years, driven by the fast expansion of new emerging applications such as the Internet of Things (IoT), social network, virtual reality and cloud computing, data center (DC) interconnect traffic over short reach links is growing rapidly. A 2016 Cisco report forecasts that in 2020 more than 80% of the global Internet traffic will be DC related, of which approximately 77% is predicted to stay within DC . Such exponential growth of data center interconnect (DCI) traffic has driven the need for high-speed and cost-effective transceivers for 40~80 km short reach applications . In these scenarios, direction detection solutions are currently more attractive than coherent solutions. Current generation switch in DC typically has 32 ports, each port utilizing a Quad Small Form-factor Pluggable 28 (QSFP28) optical transceiver. Each QPSFP28 transceiver provides a net rate of 4x25 Gb/s lanes. Higher bit rate employing pulse-amplitude modulation (PAM) multi-lane such as PAM4 8x50 Gb/s, is a potential solution for next generation 400 Gb/s transceiver. However, its spectral efficiency for C-band dense wavelength-division-multiplexing (DWDM) transmission is low and optical chromatic dispersion (CD) compensation is needed [5–8].
To further increase spectral efficiency, one potential solution is to utilize polarization-division-multiplexing (PDM) in direct detection based on a Stokes-vector receiver (SVR) [9–11]. Employing a SVR, two and three degrees of freedom (DOF) have been exploited and several high bit rate transmission demonstrations have been reported: 360 Gb/s at 20 km using PDM-2-ring/8PSK, 462 Gb/s at 1 km using PDM-PAM4 + 3PM, and 336 Gb/s at 80 km using PDM-16QAM-PAM4 [12–14]. Although the throughput, DOF and spectral efficiency have increased over the past three years, the use of SVR-based modulation is restricted by the three DOFs in the Stokes space [15–18].
Another active research topic is to employ single side band signaling in order to mitigate channel impairments and to exploit an extra DOF compared to conventional IM/DD [19, 20]. In 2016, a direct detection scheme based on Kramers Kronig (KK) relation was reported . This approach enables the full complex field reconstruction after power detection by removing signal-to-signal beat interference (SSBI) with the assumption of minimum phase condition. The KK technique has been shown to have a superior performance over the reported SSBI cancellation techniques in the literature [21–24]. Using the KK detection, several high speed transmissions have been demonstrated: 240 Gb/s/λ over 125 km fiber using a heterodyne approach at receiver, 256 Gb/s/λ over 160 km fiber aided with a digital carrier [23,24]. However, the spectral efficiency of these KK systems is limited by the single polarization operation or large guard band between the signal and carrier.
In this paper, we report the first single wavelength 400 Gb/s (a raw rate of 480 Gb/s including FEC overhead) direct detection transmission over 80 km of standard single mode fiber (SSMF) using a novel Stokes Vector Kramers Kronig (SVKK) transceiver to achieve four dimensional (4D) modulation and digital chromatic dispersion (CD) compensation. The system is realized by sending a digital tone together with the PDM signals, and using the SVKK approach to de-multiplex polarizations in the Stokes space and then retrieve complex fields. Note that the digital carrier method is used in an independent work in . Firstly, we demonstrate the feasibility of digital polarization de-multiplexing and study system performance at low baud rates (<30 GBaud) due to the lack of high speed analog-to-digital-converter (ADC) channels. It is shown that with different state-of-polarization rotations (SOP) at the receiver, the variation of received signal-to-noise ratio (SNR) is less than 0.2 dB. We also investigate the optimization of the carrier-to-signal-power-ratio (CSPR) and the guard band between the digital tone and the signal. In addition, other parameters are shown to have an impact on the KK performance including roll-off factor and oversampling ratio. Lastly, we demonstrate dual-polarization signals at high baud rates (>30 GBaud). A 480 Gb/s single carrier signal with 60 Gbaud PDM-16QAM is successfully transmitted over 80 km SSMF with a bit-error-rate (BER) below the soft-decision forward error correction (SD-FEC) threshold of 2.0x10−2 with a 20% overhead. Figure 1 shows the data rate and distance of this work relative to other reported direct detection schemes in the literature. To the best of our knowledge this work reports the highest data rate per wavelength for >1 km direct detection transmissions. Toward the completion of this paper, it came to our attention that a similar yet different polarization demultiplexing scheme that combines a Stokes vector receiver with Kramers Kronig field reconstruction was proposed and verified by simulation in parallel in .
The remainder of the manuscript is organized as follows. Section 2 presents the principle of the proposed SVKK scheme for 4D modulation direct detection. Section 3 details the experimental setup and the demonstration of the SVKK based short reach transmission using DP-16QAM modulation. Section 4 concludes the paper.
2. Principle of 4D PDM SVKK transceiver
The architecture of the proposed SVKK transceiver is illustrated in Fig. 2 with digital signal processing (DSP) blocks. At the transmitter in Fig. 2(a), complex double sideband QAM signals and are generated after pulse shaping and pre-emphasis. Then a digital carrier is inserted at the right edge with a fixed guard band. It is noted that the use of digital carrier offers complexity reduction at transmitter side (no additional optical tone). A dual-polarization (DP) IQ modulator (IQM) is used to generate 4D optical signals. At the receiver, a SVR is employed to enable polarization de-multiplexing for PDM signals in the Stokes space [26,27]. Denoting and as the real and imaginary parts of the beating between the two transmitted signals (data + carrier), respectively, and ignoring noise, the polarization rotation can be recovered asFig. 2(b). The de-rotation matrix in Eq. (1) can be obtained using either least mean square (LMS) adaptation or Stokes-orthogonal training samples [26,27].
After the and are obtained, the KK detection is applied to these power signals to recover the complex field as Eq. (2) are then low pass filtered to remove out of band noise.
The filtered and are fed into a 4x4 real multiple-input and multiple-output (MIMO) LMS filter to remove linear interference between and and to mitigate the impact of imbalance (phase, amplitude and skew) among the four transmitted components (real and imaginary of X and Y polarizations) :Figure 2(b) shows the constellations of 60 Gbaud PDM-16QAM signals before (Gaussian-like shape due to CD) and after the MIMO filter, indicating its effectiveness. The outputs of Eq. (3) are the recovered data symbols, and the BER is counted based on the hard decision of those symbols.
3. Experimental setup and results
3.1 Experimental setup
The experimental setup is depicted in Fig. 3. The output of an external cavity laser operating at 1550.12 nm with 15.5 dBm optical power was fed into a DP-IQM with 35 GHz 3-dB bandwidth. Four 8-bit DACs operating at 88 GS/s generated four electrical signals which were amplified using four discrete radio-frequency (RF) amplifiers. Each RF amplifier has 3-dB bandwidth of 50 GHz. The four outputs of the RF amplifiers were used to drive the DP-IQM. The generated optical signal was passed through a 0.4 nm optical filter to remove the residual higher order harmonic components from the DACs. The output of this filter, of which the optical spectrum is plotted in Fig. 3, was boosted by an Erbium-doped fiber amplifier (EDFA) and the optical launch power was set to be 8 dBm using a variable optical attenuator (VOA) before the signal entered a spool of 80 km SSMF. After transmission, the signal was filtered by a 0.8 nm optical filter and amplified before the SVR, which is described in Fig. 2(b), with a received power of 12 dBm. The photodiodes (PDs) and balanced PDs had 3-dB bandwidths of 40 GHz. None of the PDs hosted trans-impedance amplifiers.
The transceiver DSP blocks, which are described previously in Fig. 2, consisted of root-raised-cosine pulse shaping, digital carrier insertion with a predefined CSPR and guard band, and a MIMO filter as Eq. (3) with 61 real taps to compensate residual linear interference and CD. The CSPR, guard band and roll-off factor were optimized to achieve the best performance. The total length of the generated random bit sequence was 635,904 bits for each polarization, and the number of symbols used in the BER calculation was more than 80,000 symbols per polarization.
3.2.1 System study and demonstration of polarization de-multiplexing with 27 Gbaud signals
In the first part of this experiment, we operated the SVKK system at low baud rates (<30 GBaud) to study the impact of CSPR and demonstrate the feasibility of polarization de-multiplexing. The received signals were digitized using a real-time oscilloscope (RTO) (Keysight DSO-X 96204Q) with four 33 GHz ADC channels. The sampling rate was 80 GS/s and the resolution was 8 bits. Figure 4(a) shows the impact of CSPR on 27 Gbaud 16QAM signals in a back-to-back (B2B) scenario. We can see that the optimal CSPR depends on the guard band, and generally it increases as the guard band (≥2 GHz) gets larger. This is because the DAC has limited bandwidth and finite frequency dependent effective number of bits (ENoB), which makes higher frequency digital tones noisier. For a 1 GHz guard band, there can be additional crosstalk between the signal and the tone. With a 11.5 dB CSPR and a 4 GHz guard band, we show in Fig. 4(b) the BER versus optical-signal-to-noise-ratio (OSNR) with 0.1 nm resolution at the B2B for 27 Gbaud 16QAM and 8QAM signals. It is noteworthy to mention that the proposed digital carrier approach has high transceiver noise (due to low electrical signal power) which can lead to an increased required OSNR in B2B.
In order to perform polarization demultiplexing, we use training Stokes vector samples based on Eq. (1). In our experiment, the de-rotation process is conducted with one set of training symbols for every capture, which contains 2 million samples at a sampling rate of 80 GSa/s. In practical implementation, a periodical training or adaptive algorithm is needed to track the time-varying SOP [9, 26]. During the training period, the digital carrier is not inserted. A periodic pattern of [1, 0, 1, 1] and [0, 1, 1, i] at 0.068 Gbaud is applied for Ex and Ey, respectively, to estimate the polarization rotation matrix in Eq. (1). Similar approaches can be found in previously reported SVR systems [9, 26, 29]. In Fig. 5(a), we show the synchronized training samples of transmitted and received in two incident SOP scenarios at 80 km transmission. We can see that the training Stokes pattern is tolerant to the accumulated CD over 80 km SSMF. In this work, we use an averaging window of 128 samples centered in the middle of each time training slot to compute the de-rotation matrix in Eq. (1). In Fig. 5(b), we compare the performance of the proposed SVKK system with different states of polarization (SOP) of the 27 Gbaud 16QAM signal at the receiver. Results show that the signal-to-noise ratios (SNR) for the three SOPs, either at B2B or after 80 km, are less than 0.2 dB, which indicates the effectiveness of the SVKK polarization de-multiplexing.
In Fig. 6, we conduct a parametric study on the proposed SVKK approach with a CSPR of 11.5 dB. Figure 6(a) shows the BER versus baud rate using 8QAM, 16QAM, 32QAM and 64QAM in B2B. It indicates that 32QAM and 64QAM suffer from limited ENoB even at the low frequency regime. The effect is more severe in our approach since the digital carrier consumes significantly larger power compared to signal power. Figure 6(b) shows the impact of the roll-off factor (ROF) for 27 Gbaud 16QAM signals with a 4 GHz guard band in B2B. From this, the Nyquist-like pulse shaping, i.e. ROF close to 0, induces small penalty (~0.5 dB SNR). The explanation is that a Nyquist pulse has a higher peak-to-average-power ratio (PAPR) in time domain, which in turn may violate the minimum phase condition for the KK detection. Figure 6(c) shows the effectiveness of the LMS in term of CD compensation at various digital oversampling ratio. It demonstrates similar performance between the three methods: LMS only, LMS + CD precompensation, and LMS + CD postcompensation. It is noted that the LMS with real MIMO has to be used in the cases due to the inherent imbalance at the transmitter side. Figure 6(d) shows the recovered constellation using the conventional 2x2 butterfly complex MIMO and 4x4 real MIMO with a difference in SNR of ~1dB [30,31].
3.2.2 Demonstration of 480 Gb/s/λ transmission
Since the feasibility of the polarization de-multiplexing, which is independent of baud rate, has been verified as shown in Fig. 5, in the second part of the experiment the SOP of the incident signal was aligned to avoid polarization mixing, and the Stokes polarization de-multiplexing block in Fig. 2(b) was not employed. This was because we had to multiplex the four RTO channels to two 63 GHz bandwidth channels with a sampling rate of 160 GS/s for the detection of the power terms and at baud rates higher than 30 Gbaud. The CSPR and guard band were optimized for each baud rate. For 60 Gbaud, the CSPR was 13.5 dB and the guard band was 0.5 GHz. For 32.5 Gbaud, the CSPR was 12 dB and the guard band was 3.5 GHz. Figure 7(a) shows the BER versus baud rate for both the B2B and 80 km transmission scenarios. 32.5 Gbaud (raw bit rate 260 Gb/s) and 60 Gbaud (raw bit rate 480 Gb/s) PDM-16QAM signals were transmitted over 80 km with a BER below the hard-decision forward error correction (HD-FEC) threshold of 3.8x10−3 and SD-FEC threshold of 2.0x10−2, respectively. For the latter, excluding 20% FEC overhead, the net data rate is 400 Gb/s. To the best of our knowledge, this is the first experimental demonstration of single wavelength direct detection 400G system over 80 km fiber. The oversampling ratio is an important parameter in the KK detection . Figure 7(b) illustrates the impact of the digital oversampling ratio for signals with various baud rates prior to the KK operation. An important observation is that for any baud rate, the SNR always suffers from a large penalty at 2 samples per symbol as already pointed out in . This can be explained by Eq. (2), where a sufficient oversampling ratio (>3 samples/symbol) allows for an accurate waveform generation from nonlinear operations, i.e. square root and logarithm.
We have proposed a SVKK system using a digital carrier to enable 4D modulation direct detection over dispersion-unmanaged C-band transmission. The feasibility of digital polarization de-multiplexing was experimentally demonstrated, and the impact of system parameters and configurations including carrier-to-signal-power ratio, guard band of the digital carrier, oversampling ratio and real MIMO was experimentally investigated. With 60 Gbaud PDM-16QAM signals, we reported the first single wavelength 480 Gb/s (net bit rate 400 Gb/s) transmission over 80km fiber in C-band.
References and links
1. M. Y. S. Sowailem, T. M. Hoang, M. Morsy-Osman, M. Chagnon, D. Patel, S. Paquet, C. Paquet, I. Woods, O. Liboiron-Ladouceur, and D. V. Plant, “400G single carrier 500km transmission with an InP dual polarization IQ modulator,” IEEE Photonics Technol. Lett. 28(11), 1213–1216 (2016). [CrossRef]
2. K. Roberts, Q. Zhuge, I. Monga, S. Gareau, and C. Laperle, “Beyond 100 Gb/s: capacity, flexibility, and network optimization,” IEEE J. Opt. Commun. Netw. 9(4), C12–C24 (2017). [CrossRef]
3. Cisco Global Cloud Index, November (2016). https://www.cisco.com/c/dam/en/us/solutions/collateral/service-provider/global-cloud-index-gci/white-paper-c11-738085.pdf
4. K. P. Zhong, X. Zhou, Y. Wang, T. Gui, Y. Yang, J. Yuan, L. Wang, W. Chen, L. Zeng, C. Yu, A. P. T. Lau, and C. Lu, “Recent advances in short reach systems,” in Optical Fiber Communication Conference (Optical Society of America, 2017), paper Tu2D.7. [CrossRef]
5. IEEE 802.3, [online], http://www.ieee802.org/3.
6. K. P. Zhong, X. Zhou, Y. Gao, W. Chen, J. Man, L. Zeng, A. P. T. Lau, and C. Lu, “140 Gbit/s 20km transmission PAM-4 signal at 1.3mm for short reach communications,” IEEE Photonics Technol. Lett. 27(16), 1757–1760 (2015). [CrossRef]
7. M. Birk, L. E. Nelson, G. Zhang, C. Cole, C. Yu, M. Akashi, K. Hiramoto, X. Fu, P. Brooks, A. Schubert, T. Baldwin, R. Luking, and G. Pepper, “First 400GBASE-LR8 interoperability using CFP8 modules,” in Optical Fiber Communication Conference (Optical Society of America, 2017), paper Th5B.7.
8. N. Eiselt, J. Wei, H. Griesser, A. Dochhan, M. H. Eiselt, J. Elbers, J. J. V. Olmos, and I. T. Monroy, “Evaluation of real-time 8 × 56.25 Gb/s (400G) PAM-4 for inter-data center application over 80 km of SSMF at 1550 nm,” J. Lightwave Technol. 35(4), 955–962 (2017). [CrossRef]
10. M. Morsy-Osman, M. Chagnon, M. Poulin, S. Lessard, and D. V. Plant, “224-Gb/s 10-km transmission of PDM PAM-4 at 1.3 μm using a single intensity modulated laser and a direct detection MIMO DSP-based receiver,” J. Lightwave Technol. 33(7), 1380–1387 (2015). [CrossRef]
11. K. Kikuchi, “Simple and efficient algorithm for polarization tracking and demultiplexing in dual-polarization IM/DD systems,” in Optical Fiber Communication Conference (Optical Society of America, 2015), paper Th1E.3. [CrossRef]
12. M. Morsy-Osman, M. Chagnon, and D. V. Plant, “Four-dimensional modulation and Stokes direct detection of polarization division multiplexed intensities, inter polarization phase and inter polarization differential phase,” J. Lightwave Technol. 34(7), 1585–1592 (2016). [CrossRef]
13. M. Chagnon and D. V. Plant, “504 and 462 Gb/s direct detect transceiver for single carrier short-reach data center applications,” in Optical Fiber Communication Conference (Optical Society of America, 2017), paper W3B.2. [CrossRef]
14. T. Hoang, M. Sowailem, Q. Zhuge, M. Osman, A. Samani, C. Paquet, S. Paquet, I. Woods, and D. V. Plant, “Enabling high-capacity long reach direct detection transmission with QAM-PAM Stokes vector modulation,” J. Lightwave Technol. (to be published).
15. P. Dong, X. Chen, K. Kim, S. Chandrasekhar, Y.-K. Chen, and J. H. Sinsky, “128-Gb/s 100-km transmission with direct detection using silicon photonic Stokes vector receiver and I/Q modulator,” Opt. Express 24(13), 14208–14214 (2016). [CrossRef] [PubMed]
16. M. Y. S. Sowailem, T. M. Hoang, M. Chagnon, M. Morsy-Osman, M. Qiu, S. Paquet, C. Paquet, I. Woods, O. Liboiron-Ladouceur, and D. V. Plant, “100G and 200G single carrier transmission over 2880 and 320 km using an InP IQ modulator and Stokes vector receiver,” Opt. Express 24(26), 30485–30493 (2016). [CrossRef] [PubMed]
17. A. Li, Z. Li, Y. Wen, W. Peng, Y. Cui, and Y. Bai, “192-Gb/s 160-km transmission of carrier-assisted dual-polarization signal with Stokes vector direct detection,” in Optical Fiber Communication Conference (Optical Society of America, 2016), paper W1A.2. [CrossRef]
18. D. Che, Q. Hu, and W. Shieh, “Linearization of direct detection optical channels using self-coherent subsystems,” J. Lightwave Technol. 34(2), 516–524 (2016). [CrossRef]
19. S. Randel, D. Pilori, S. Chandrasekhar, G. Raybon, and P. Winzer, “Linearization of direct detection optical channels using self-coherent subsystems,” in Proceedings of European Conference on Optical Communication (ECOC) (Institute of Electrical and Electronics Engineers, 2015), paper Mo.4.5.2.
20. J. C. Cartledge and A. S. Karar, “100 Gb/s intensity modulation and direct detection,” J. Lightwave Technol. 32(16), 2809–2814 (2014). [CrossRef]
21. A. Mecozzi, C. Antonelli, and M. Shtaif, “Kramers-Kronig coherent receiver,” Optica 3(11), 1220–1227 (2016). [CrossRef]
22. Z. Li, M. S. Erkilinc, K. Shi, E. Sillekens, L. Galdino, B. C. Thomsen, P. Bayvel, and R. I. Killey, “SSBI mitigation and Kramers-Kronig scheme in single-sideband direct-detection transmission with receiver-based electronic dispersion compensation,” J. Lightwave Technol. 35(10), 1887–1893 (2017). [CrossRef]
23. S. T. Le, K. Schuh, M. Chagnon, F. Buchali, R. Dischler, V. Aref, H. Buelow, and K. Engenhardt, “8x256 Gbps virtual-carrier assisted WDM direct-detection transmission over a single span of 200km,” in Proceedings of European Conference on Optical Communication (ECOC) (Institute of Electrical and Electronics Engineers, 2017), paper Th.PDP.B1.
24. X. Chen, C. Antonelli, S. Chandrasekhar, G. Raybon, A. Mecozzi, M. Shtaif, and P. Winzer, “4x240 Gb/s dense WDM and PDM Kramers-Kronig detection with 125-km SSMF transmission,” in Proceedings of European Conference on Optical Communication (ECOC) (Institute of Electrical and Electronics Engineers, 2017), paper W.2.D.4.
25. C. Antonelli, A. Mecozzi, M. Shtaif, X. Chen, S. Chanrasekhar, and P. Winzer, “Polarization multiplexing with the Kramers Kronig receiver,” J. Lightwave Technol. (to be published).
26. M. Chagnon, M. Osman, D. Patel, V. Veerasubramanian, A. Samani, and D. V. Plant, “Digital signal processing for dual-polarization intensity and inter-polarization phase modulation formats using Stokes detection,” J. Lightwave Technol. 34(1), 188–195 (2016). [CrossRef]
27. D. Che and W. Shieh, “Polarization demultiplexing for Stokes vector direct detection,” J. Lightwave Technol. 34(2), 754–760 (2016). [CrossRef]
28. M. Faruk and K. Kikuchi, “Compensation for in-phase/quadrature imbalance in coherent-receiver front end for optical quadrature amplitude modulation,” IEEE Photonics J. 5(2), 7800110 (2013). [CrossRef]
29. M. Sowailem, T. Hoang, M. Osman, M. Chagnon, M. Qiu, S. Paquet, C. Paquet, I. Woods, O. Liboiron-Ladouceur, and D. V. Plant, “Impact of chromatics dispersion compensation in single carrier two-dimensional stokes vector direct detection,” IEEE Photonics J. 9(4), 7203110 (2017). [CrossRef]
31. S. Tsukamoto, D.-S. Ly-Gagnon, K. Katoh, and K. Kikuchi, “Coherent demodulation of 40-Gbit/s polarization-multiplexed QPSK signals with 16-Ghz spacing after 200-km transmission,” in Optical Fiber Communication Conference (Optical Society of America, 2015), paper PDP-29.