Open Access
Add to BibSonomyAbstract
In this paper, we analyze the stimulated Raman scattering (SRS) effect between multiple TWDM downstream wavelengths in L band and multiple fronthaul wavelength channels in C band theoretically, and investigate its impact on fronthaul signals experimentally for the first time. The impact includes two aspects, one is SRS-induced power depletion and the other is the eye diagram distortion caused by the nonlinear Raman crosstalk. Experimental results show that, up to 1-dB power depletion would be introduced to each fronthaul wavelength channel with the launch power per TWDM channel varying from 8-dBm to 15-dBm. In addition, due to the “1” level broadening of the non-return-to-zero on-off keying (NRZ-OOK) eye diagram, the receiving sensitivity with 20-km standard single mode fiber (SSMF) transmission is decreased by ~1-dB. Therefore, a total of ~2-dB power penalty would be imposed on fronthaul signals when co-transmitting with TWDM downstream signals in the same feeder fiber.
© 2015 Optical Society of America
1. Introduction
Recently, with the maturing of fourth-generation (4G) standardization and the ongoing worldwide deployment of 4G cellular networks, research on fifth-generation (5G) communication technologies have been started in both the academic and industrial communities [1]. It is envisioned to support 1000 times higher mobile data volume per area, 10 to 100 times higher number of connected devices and 10 ~100 times higher user data rate [2] through the promising centralized or cloud radio access network (C-RAN) architecture [1,3]. In C-RAN, remote radio heads (RRHs) are connected through optical fiber to the centralized baseband units (BBUs) using the point-to-point wavelength division multiplexing (PtP WDM) overlay, which is recommended in the next generation passive optical network stage2 (NG-PON2) standard G.989.1 as fronthaul [4,5]. The responsibility of the fronthaul link is to transmit IQ samples of the complex baseband radio signal through a standard interface, for example Common Public Radio Interface (CPRI) [6]. Nevertheless, due to the usage of massive multiple-input multiple-output (MIMO) antenna technology in 5G cellular system, hundreds of gigabits backhaul capacity in ultra-dense cell networks is needed. It results in lots of fronthaul wavelengths transmission in the optical fiber. At the same time, owing to the characteristic of 5G communications, the mobile fronthaul has stringent requirement on the link latency, synchronization, jitter and bit error ratio (BER) etc [7, 8].
However, except for the fronthaul wavelengths in C band, other downstream wavelengths from gigabit-capable passive optical network (GPON), 10-gigabit-capable passive optical network (XG-PON), time and wavelength division multiplexed passive optical network (TWDM-PON) [4,9] and RF-video may also exist in the same feeder fiber due to the system overlay. These downstream channels are assigned in various wavebands, including 1490-nm, 1577-nm, 1600-nm and 1550-nm. Consequently, on this full coexistence scenario, the fiber nonlinearity especially the stimulated Raman scattering (SRS) may occur in the feeder fiber, which would deteriorate the performance of mobile fronthaul links. In this paper, only downstream transmission is considered rather than the upstream. Because the optical power in the feeder fiber during the upstream transmission is much lower than the downstream owing to high loss of the PON power splitters. As we all know that, SRS causes power transfer from the Pump signal to the Stokes signal, which would induce two impacts on the Pump signal. One is the SRS-caused power depletion and the other is the crosstalk noise on Pump signal due to the Raman nonlinearity. For instance, on the coexistence of GPON operating at 1490-nm and TWDM-PON operating at 1600-nm, 4-dB SRS-caused power depletion would be experienced on GPON for the case of 8 TWDM wavelengths [10,11]. As for the suppression of nonlinear Raman crosstalk on RF-video signal, many researchers have proposed corresponding techniques, for example simple RF filtering in [12,13] and delay modulation in [14,15]. We also proposed dicode coding to reduce this SRS-caused nonlinear crosstalk in [16]. All these techniques are intended to improve the signal-to-noise-ratio (SNR) of analogy RF-video signal. Similarly, for digital signal transmission, the eye opening of non-return-to-zero on-off keying (NRZ-OOK) signal would also be degraded by the SRS-induced nonlinear crosstalk [17]. And this impact on NRZ-OOK signal of fronthaul wavelengths under multiple TWDM wavelengths has never been experimentally investigated yet.
In this paper, we focus on the SRS effect between multiple TWDM downstream wavelengths in L band and multiple fronthaul wavelengths in C band, and setup an experiment to investigate its impact on fronthaul signals for the first time. Experimental results indicate that, up to 1-dB signal power would be depleted on each fronthaul wavelength channel when the launch power of each TWDM channel varies from 8-dBm to 15-dBm. Besides, the receiving sensitivity with 20-km fiber transmission is decreased by ~1-dB owing to the “1” level broadening of the NRZ-OOK eye diagram. As a result, the SRS effect would cause ~2-dB power penalty on the fronthaul signals under four TWDM downstream wavelengths and four fronthaul wavelengths coexisting scenarios.
2. Theoretical analysis of the stimulated Raman scattering between Fronthaul and TWDM wavelengths
Figure 1 shows the downstream transmission with TWDM and fronthaul wavelength channels. In the optical line terminal (OLT), the multiplexed fronthaul and TWDM wavelength channels are combined through a co-existence element [4]. The whole wavelength channels are depicted in the inset (i) of Fig. 1. In this system, the SRS interaction between fronthaul and TWDM wavelength channels can be described as following [18].
where PT is the power level per TWDM channel and PF is the power level of each fronthaul wavelength channel. The group velocity of TWDM and fronthaul channels is vT and vF, respectively. g is the standard Raman gain coefficient (gR) divided by the fiber effective area (Aeff), which can be expressed as g = gR/Aeff, and α is the fiber linear loss. Firstly, we only consider the power depletion on fronthaul wavelengths. Hence, the SRS term (gPFPT) resulting in the nonlinear power gain on TWDM channels can be ignored. Regardless of the time-related term on the left hand side of Eqs. (1) and (2), the SRS-caused power depletion ΔP under N TWDM channels can be solved as follows [10]. where PT(0) is the launch power per TWDM wavelength channel. Leff is the effective fiber length, which is determined by the fiber length L and the fiber linear loss α as expressed in Eq. (4). Then, we further consider the nonlinear Raman crosstalk on fronthaul wavelength channels. It can be defined as the ratio of the power of SRS-induced crosstalk on fronthaul wavelengths to the digital modulation signal power on TWDM wavelength channels [19], which can be expressed as following.Where d is the group velocity mismatch between TWDM and fronthaul wavelengths, which can be approximated as. The parameter D represents the dispersion coefficient (ps/nm/km), and the wavelength of TWDM and fronthaul signals isandrespectively. N is the number of TWDM wavelength channels, L is the fiber length and PT(0) is the launch power per TWDM wavelength channel.
Fig. 1 The downstream transmission with TWDM-PON and fronthaul signals. WDM: wavelength division multiplexing; ONU: optical network unit; RRH: remote radio head; inset (i) is the whole downstream wavelengths.
Based on the analyzed results, we calculate the SRS-caused power depletion on fronthaul wavelengths and the nonlinear Raman crosstalk in case of 20-km fiber transmission. Generally, the number of wavelengths in TWDM-PON is 4 or 8 according to the TWDM standard G.989.1 [4]. However, considering the rapid development of PON technologies, the number of TWDM wavelength channels in the calculation is varied from 4 to 32 [21]. Figure 2 and Fig. 3 show the calculated SRS-caused power depletion and the nonlinear Raman crosstalk, respectively. It can be clearly seen that, the SRS-caused power depletion increases with the increase of the launch power per TWDM channel. And the power depletion on one fronthaul wavelength is the sum of the power transfer to each TWDM wavelength channel, so the more the number of TWDM downstream wavelengths is, the more power depletion would be experienced on each fronthaul wavelength channel. Similarly, the nonlinear Raman crosstalk also increases as the number of TWDM wavelengths increases. And due to the low-pass characteristic of the nonlinear Raman crosstalk, it has more severe influence on the low frequency region than the high frequency as depicted in Fig. 3. In order to clarify the impact on NRZ-OOK signal quality, we conduct an experiment. In the following section, we will show the experimental setup and the results in detail.
Fig. 2 The calculated SRS-caused power depletion on fronthaul wavelengths at different number of TWDM downstream wavelengths (NTWDM = 4, 8, 16 or 32).
Fig. 3 The calculated SRS-caused nonlinear crosstalk on fronthaul wavelengths at 13-dBm launch power per TWDM channel (NTWDM = 4, 8, 16 or 32).
3. Experimental setup
Figure 4 shows the experimental setup for investigation of the SRS effect between multiple fronthaul wavelengths and multiple downstream wavelengths in TWDM-PON. In this experiment, eight distributed feedback (DFB) lasers are utilized to generate four fronthaul wavelengths (λ1-λ4: 1529.1nm, 1529.9nm, 1530.7nm and 1531.5nm) and four TWDM downstream wavelengths (λ5-λ8: 1597.1nm, 1597.9nm, 1598.7nm and 1599.5nm), which is shown in the inset (i) of Fig. 4. The output power of each DFB laser is ~10-dBm and each DFB laser followed by a polarization controller (PC) is multiplexed through a 100-GHz arrayed waveguide grating (AWG). The PCs are used to adjust the polarization of each TWDM or fronthaul wavelength channel to minimize the optical loss caused by the polarization-sensitive MZM modulators. Once the MZM-induced optical loss for each TWDM/fronthaul wavelength is minimized, the polarization state of all TWDM or fronthaul wavelengths is assumed to be aligned independently. And the extra PCs placed between the coupler and EDFA are utilized to maximum the crosstalk during the Raman noise measurement. Then, the multiplexed fronthaul or TWDM downstream wavelengths are injected into a standard single-electrode Mach-Zehnder Modulator (MZM) for modulation. The extinction ratio of the signal on TWDM and fronthaul channels is more than 10-dB. The MZM is biased at the quadrature point and driven by the signal generated from the pulse pattern generator (PPG). According to CPRI specifications [6], the signal bit rate modulated on fronthaul wavelengths is varied from 614.4-Mbit/s to 10137.6-Mbit/s while the bit rate modulated on TWDM wavelengths is 10-Gbit/s. In order to study the nonlinear Raman crosstalk under different launch powers, we use a C/L band Erbium doped fiber amplifier (EDFA) to adjust the launch power of fronthaul/TWDM wavelength channels. Then, the fronthaul wavelengths are combined with TWDM downstream signals through a 50/50 coupler before launched into the 20-km SSMF for transmission. At the receiver, a tunable optical filter (TOF) is utilized to select the fronthaul wavelengths and a PIN is employed to detect the fronthaul signals for the further analysis. The BER performance and the nonlinear Raman crosstalk are measured through a BER tester and a spectrum analyzer, respectively. The resolution bandwidth (RBW) and video bandwidth (VBW) of the spectrum analyzer is 30-kHz in the measurement.
Fig. 4 Experimental setup for measuring the SRS-caused power depletion and the nonlinear Raman crosstalk using four fronthaul wavelengths (λ1-λ4) and four TWDM downstream wavelengths (λ5-λ8); Inset (i) is the whole wavelengths in the experiment. PPG: pulse pattern generator; AWG: arrayed waveguide grating; MZM: Mach-Zehnder Modulator; PC: polarization controller; TOF: tunable optical filter.
Note that, due to lack of sufficient MZMs in our lab, the four fronthaul or TWDM wavelengths are modulated by the same data in our experiment. However, the experimental results are not influenced because the influence of the Raman noise on fronthaul signals is associated with the signal power spectrum density (PSD) [20] and the bit pattern interaction between TWDM and fronthaul signals [22]. In other words, the influence of the Raman noise on fronthaul signals only happens when there is a mark in both TWDM and fronthaul channels. If a space appears in either channel, no influence would occur. It means that, if all NRZ signals are exactly identical for each of the two groups (TWDM and fronthaul wavelengths), then the influence on fronthaul signals caused by the Raman noise would be the most serious. Conversely, there is no influence if the NRZ streams on TWDM and fronthaul channels are totally conversed. In our experiment, the PRBS with the word length of 231-1 is modulated on fronthaul channels while the TWDM channels use the 223-1 PRBS stream, which results in the irrelevant property between TWDM and fronthaul NRZ streams.
4. The experimental results and analysis
Figure 5 shows the measured SRS-caused power depletion under four fronthaul channels and four TWDM wavelengths. The launch power per fronthaul channel is 9-dBm and the launch power of each TWDM channel is varied from 8-dBm to 15-dBm. It can be seen that, the power depletion increases as the launch power per TWDM channel increases. And this increasing trend is almost consistent with the calculated result (dash line in Fig. 5) except for little differences on the exact power depletion values. This is because that, the polarization of all TWDM and fronthaul channels is assumed to be aligned in the calculation. But in the experiment, the polarization of all wavelengths is not completely in the same direction although we adjusted it to maximize the crosstalk. At 15-dBm launch power per TWDM channel, ~1-dB power depletion is induced on fronthaul channels. In addition to the power depletion, the nonlinear Raman crosstalk is also investigated. Figure 6 shows the measured nonlinear Raman crosstalk at different launch power per TWDM channel. It is normalized to the TWDM signal’s power on DC to 200-MHz frequency range. It is obvious that, the crosstalk on low-frequency region is larger than that of high-frequency region, which is in agreement with the calculation in section II. And as the launch power per TWDM channel increases from 11-dBm to 15-dBm, the crosstalk on DC to 200-MHz frequency region also increases slowly. Because when increasing the launch power per TWDM channel, more power on TWDM channels are transferred to fronthaul wavelengths. It should be noted that, in the measurement, the fronthaul wavelengths are not modulated in order to make sure that the measured spectrum at the receiver only includes the nonlinear Raman crosstalk.
Fig. 5 Measured SRS-caused power depletion on each fronthaul channel (λ1-λ4) under four TWDM downstream wavelength channels (NTWDM = 4).
Fig. 6 Measured nonlinear Raman crosstalk on fronthaul wavelength channels caused by four TWDM downstream wavelength channels (NTWDM = 4).
Then, we modulate four fronthaul channels using NRZ signals and investigate the Raman crosstalk induced signal distortion on fronthaul channels with four TWDM wavelengths coexist. The data rate of TWDM channels is fixed at 10-Gbit/s while the data rate of fronthaul channels is varied from 1228-Mbit/s to 10137-Mbit/s. The launch power per TWDM channel is 13-dBm and the launch power of each fronthaul channel is 9-dBm. We measured thespectrum and eye diagrams of the fronthaul signal as depicted in Fig. 7. The original signal spectrum and eye diagrams with no TWDM coexistence are also depicted for comparison. Note that as the crosstalk mainly overlays on the low frequency region, we zoomed the figure and depicted only the 0~100MHz region of the signal spectrums. The three spectrums in Fig. 7 are all measured under modulated fronthaul channels instead of CW channels. For NRZ signals with different bit rates, the received power on DC to 100-MHz frequency range would be different. It can be seen from the red curves in Figs. 7(a)-7(c) that, the lower the bit rate is, the higher the power would be, which results in higher signal-to-noise-ratio (SNR) on the low frequency range. Therefore, once the nonlinear Raman noise is superimposed on the fronthaul channels, the lower bit rate signal would show a higher tolerance. For example, at the data rate of 10137-Mbit/s shown in Fig. 7(a), the power with SRS-induced nonlinear crosstalk is ~3-dB larger than that without SRS-caused nonlinear crosstalk in DC to 25-MHz frequency region. Nevertheless, when decreasing the data rate to 1228-Mbit/s, the difference between with and without nonlinear Raman crosstalk situations becomes less obvious. The original and distorted eye diagrams are also shown in the inset (i) ~(vi) of Fig. 7. Due to the existence of low-pass nonlinear Raman crosstalk, the “1” level of the eye diagram is all broadened compared to the original one, which result in a lower extinction ratio (ER) and worse sensitivity, especially for the high data rate cases.
Fig. 7 Measured DC to 100-MHz spectrum and eye diagrams (i ~vi) at the data rate of (a): 10137-Mbit/s; (b): 4915-Mbit/s; (c): 1228-Mbit/s.
Finally, the BER performance of fronthaul channels is measured in case of 20-km SSMF transmission. Figure 8 depicts the measured BER curves of different bit-rate signal when four fronthaul channels and four TWDM wavelengths are coexisted. During this measurement, the same optical receiver (New focus: model 1554-B) with 3-dB bandwidth of 12-GHz is used for all data rates. The launch power per TWDM channel is 13-dBm. At the data rate of 10137-Mbit/s, there is ~1-dB power difference on the receiving sensitivity between with and without the crosstalk. As for the lower bit-rate signal, the receiving sensitivity between with and without the nonlinear Raman crosstalk is almost the same. Because the low bit-rate NRZ-OOK signal has higher tolerance to the nonlinear Raman crosstalk as illustrated before. Note that, in this experiment, only four TWDM wavelength channels and four fronthaul wavelengths are considered. Whereas, in the future access network such as 5G cellular system, more fronthaul wavelengths and more TWDM wavelength channels would be co-transmitted in the same feeder fiber. In this situation, the nonlinear Raman crosstalk would impose more severe influence on the receiving sensitivity of fronthaul signals.
5. Conclusions
In this paper, the nonlinear Raman scattering between four fronthaul wavelengths in C band and four TWDM downstream wavelengths in L band is investigated experimentally for the first time. And we also theoretically analyze its impact on fronthaul wavelengths including SRS-caused power depletion and the eye diagram distortion. Through the experiment, we find that when the launch power per TWDM channel varies from 8-dBm to 15-dBm, 0.2-dB~1-dB signal power would be depleted on each fronthaul channel due to the SRS-caused power transfer. Moreover, due to the “1” level broadening of the NRZ-OOK eye diagram, the receiving sensitivity of 10137-Mbit/s fronthaul signal would be decreased by ~1-dB in case of 20-km SSMF transmission. Hence, a total of ~2-dB power penalty would be imposed on fronthaul signals when co-transmitting with four TWDM downstream signals in the same feeder fiber.
Acknowledgments
The work was jointly supported by the NSFC (61431009, 61221001 and 61371082), the 863 Program and the 973 Program (2012CB315602).
References and links
1. I. Chih-Lin, C. Rowell, S. Han, Z. Xu, G. Li, and Z. Pan, “Toward green and soft: a 5G perspective,” IEEE Commun. Mag. 51(2), 90–97 (2013).
2. A. Osseiran, F. Boccardi, V. Braun, K. Kusume, P. Marsch, M. Maternia, O. Queseth, M. Schellmann, H. Schotten, H. Taoka, H. Tullberg, M. A. Uusitalo, B. Timus, and M. Fallgren, “Scenarios for 5G mobile and wireless communications: the vision of the METIS project,” IEEE Commun. Mag. 52(5), 26–35 (2014). [CrossRef]
3. China Mobile Research Institute, “C-RAN the road towards green RAN,” White Paper Version 2.5 (Oct, 2011) http://labs.chinamobile.com/cran/wp-content/uploads/CRAN_white_paper_v2_5_EN%281%29.pdf
4. 40 gigabit-capable passive optical networks (NG-PON2): General requirements, ITU-T Recommendation G.989.1, 2013. http://www.itu.int/rec/T-REC-G.989.1-201303-I.
5. K. Grobe, “Next-generation access/backhaul based on ITU G.989, NG-PON2,” in Proc. of Photonic Networks 2014.
6. Common Public Radio Interface (CPRI), Interface Specification v6.1 (Jul, 2014) http://www.cpri.info/downloads/CPRI_v_6_1_2014-07-01.pdf
7. T. A. Diallo, A. Pizzinat, P. Chancloul, F. Salioul, F. Deletrel, and C. Aupetit-Berthelemot, “Jitter impact on mobile fronthaul links,” in Proc. OFC 2014, paper W2A.41.
8. T. Pfeiffer and F. Schaich, “Optical architectures for mobile back- and fron-thauling,” presented at the OFC/NFOEC Wireless Backhauling Workshop, Los Angeles, CA, USA, Mar. 2012.
9. 40-Gigabit-capable passive optical networks (NG-PON2): physical media dependent (PMD) layer specification, ITU-T Recommendation G.989.2, 2014.
10. R. Gaudino, V. Curri, and S. Capriata, “Propagation impairments due to Raman effect on the coexistence of GPON, XG-PON, RF-video and TWDM-PON,” in Proc. of ECOC2013, P.6.1206. [CrossRef]
11. V. Curri, S. Capriata, and R. Gaudino, “Outage probability due to Stimulated Raman Scattering in GPON and TWDM-PON coexistence,” in Proc. OFC 2014, paper M3I.2. [CrossRef]
12. A. Tanaka, N. Cvijetic, and T. Wang, “Beyond 5dB nonlinear Raman crosstalk reduction via PSD control of 10Gb/s OOK in RF-Video coexistence scenarios for Next-Generation PON,” in Proc. OFC 2014, paper M3I.3. [CrossRef]
13. D. Piehler, “Minimising nonlinear Raman crosstalk in future network overlays on legacy passive optical networks,” Electron. Lett. 50(9), 687–688 (2014). [CrossRef]
14. N. Cheng, M. Zhou, K. Litvin, and F. J. Effenberger, “Delay modulation for TWDM PONs,” in Proc. OFC 2014, paper W1D.3.
15. N. Cheng, M. Zhou, and F. J. Effenberger, “10 Gbit/s delay modulation using a directly modulated DFB laser for a TWDM PON with converged services [Invited],” J. Opt. Commun. Netw. 7(1), A87–A96 (2015). [CrossRef]
16. J. Li, M. Bi, H. He, and W. Hu, “Suppression of SRS induced crosstalk in RF-video overlay TWDM-PON system using dicode coding,” Opt. Express 22(18), 21192–21198 (2014). [CrossRef] [PubMed]
17. S. Schollmann, J. Leibrich, C. Wree, and W. Rosenkranz, “Impact of SRS-induced crosstalk for different modulation formats in WDM systems,” in Proc. OFC2004.
18. G. P. Agrawal, Nonlinear Fiber Optics (Academic, 1989), Chap.8.
19. M. R. Phillips and D. M. Ott, “Crosstalk due to optical fiber nonlinearities in WDM CATV Lightwave systems,” J. Lightwave Technol. 17(10), 1782–1792 (1999). [CrossRef]
20. A. Shahpari, J. D. Reis, S. Ziaie, R. Ferreira, M. Lima, A. N. Pinto, and A. Teixeira, “Multi system Next-generation PONs impact on video overlay,” in Proc. of ECOC2013, Tu.3.F.3. [CrossRef]
21. S. Smolorz, E. Gottwald, H. Rohde, D. Smith, and A. Poustie, “Demonstration of a coherent UDWDM-PON with real-time processing,” in Proc. of OFC2011, PDPD4. [CrossRef]
22. A. R. Chraplyvy, “Limitations on lightwave communications imposed by optical-fiber nonlinearities,” J. Lightwave Technol. 8(10), 1548–1557 (1990). [CrossRef]
