Open Access
Add to BibSonomy
Abstract
A disaster-resistant, point-to-multipoint network configuration that connects distributed optical terminals to a centralized optical terminal, through an exclusive fiber link, is required in edge computing strategies such as 5G, IoT applications, ultrareliable and low latency applications, high-performance content delivery, artificial intelligence applications, and hyperconverged infrastructure. Therefore, we introduce a shared link orientation that fully connects the distributed optical terminal side to the centralized optical terminal side represented by the conventional optical access network. We also propose a method that switches a fully-coupled or half-split network to the centralized optical terminal on the optical switch aggregation station side. We report a verification experiment on the proposed method to confirm the operation of wavelength conversion/optical path switching when a link failure occurs, achieving < 2 dB penalty with small back reflection in wavelength collision-free secondary link transmissions with primary link failure. The unique point of this research is that the centralized optical terminal can easily change from the link-occupancy-oriented network to the link-sharing-oriented network. This work will lead to development of a novel disaster-resistant network configuration for wavelength division multiplexing-based coherent passive optical network systems.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Optical access networks (OANs) that support next-generation wireless communications after 5G require 100-Gbit/s class high-capacity communications [1]. The application of digital coherent technology is under investigation to accommodate efficiently transmission capacities of 100 Gb/s/λ using a single transceiver [2–5]. Due to an increase in the transmission capacity per wavelength or per optical transmitter (Tx)/receiver (Rx) in OANs, the application of a digital coherent scheme is being studied [6–8]. In addition, since the requirements for connecting distributed stations in optical lines arranged around each wireless base station to a centralized station specify a short delay, the application of wavelength division multiplexing (WDM) that yields excellent short-delay access is under investigation, instead of time-division multiplexing that was introduced in conventional home fiber optic technology [9–12].
The current network (NW) configuration in OANs has a 1:N point-to-multipoint (P-MP) configuration comprising a centralized station and N distributed stations. The mainstream configuration is a passive optical network (PON) in which passive optical devices such as optical couplers (CPLs) are placed in remote nodes. The conventional OAN popularized by fiber-to-the-home technology has a cost-effective link-sharing-oriented configuration that efficiently integrates one optical fiber. However OANs for the mobile fronthaul, which are being considered to consolidate wireless base stations using optical fibers, are link-occupancy-oriented in which the optical Tx/Rx in the centralized station and the optical Txs/Rxs in distributed stations are connected by a single optical fiber. Comparing the link-sharing-oriented type and the link-occupancy-oriented type from the viewpoint of disaster resistance, the mobile fronthaul configuration, which is classified as a link- occupancy-oriented type, exhibits low loss because independent links connect the transceivers in the distributed stations and the centralized station, so that branch loss due to the optical CPL does not occur. However, there are two issues in terms of cost. First, as the number of occupied links increases, the number of optical fibers to be deployed tends to increase, increasing the cost and posing an economic problem. In addition, there is a problem in that the number of additional optical fibers will increase when the redundancy of the optical fibers is taken into consideration because it is impossible to deal with a case where a link failure occurs in any of the optical fibers in the event of a disaster.
A fully coupled PON configuration, classified as a link-sharing-oriented type, exhibits low cost because single sharing link connects the transceivers in the distributed stations and the centralized station. However, the fully coupled PON configuration has a drawback: the branch loss is significant because optical CPLs have multiple branches/couplings. In addition, since the optical fiber links in the access section are concentrated in one place, it is vulnerable to link failures in the event of a disaster. Several protection methods were studied by adding redundant optical fibers to standardized disaster-resistant PON configurations [13–17]. For example, there is a configuration in which redundant fibers are added while maintaining branch loss by utilizing unused input/output terminals of the 2×2 optical CPLs at a remote node [18].
A PON configuration that is half-divided by incorporating the concept of a link occupancy-oriented type into a PON that is a link-sharing-oriented type increases the number of deployed optical fibers compared to the conventional PON but can reduce branch loss. The number of deployed optical fibers can be reduced, although the branch loss increases compared to the fully occupied link configuration. Therefore, the half-divided PON configuration is an intermediate approach for the two trade-off indicators of cost and branch loss.
A shared protection method that applies a single shared link to a half-split PON has not yet been thoroughly investigated. In the shared protection method for a half-split PON, to avoid the influence of wavelength collision and reflection between PONs, a function of the wavelength changing for a path group in each PON is required when the primary link fails in single-core bidirectional transmission in the same wavelength band. Therefore, by arranging a tunable light source in the optical Tx/Rx arranged in each PON, a colorless function that variably assigns an arbitrary wavelength to the optical network unit (ONU) on the terminal side is required. In the recently reported coherent WDM-PON system, an optical CPL is used to connect the ONU to the trunk span fiber on the station side optical line terminal (OLT) unit, and a colorless function is actualized using a tunable light source on the optical Tx/Rx side of the ONU and OLT [19,20]. However, there is still a problem related to the colorless function for wavelength allocation that must be examined that pertains to the influence of reflection due to link failure.
In this paper, as the target of a half-split PON structure, we propose a coherent PON that arbitrarily switches between a low-loss occupied link that implements a bypass function between the ONU and OLT and a shared link for disaster response that has a backup function. We also compare the characteristics of a fully coupled PON and half-split PON with and without link failure. When there is no failure of the primary link, the entire system can operate at low power consumption using the low-loss main link while driving a small number of light sources on the OLT side. When the primary link fails, communications are reconnected using the secondary link, and the influence of end-face reflection due to the link failure is avoided using the wavelengths allocated to the links between the PONs. A 100-Gb/s/λ/path × 4 ONU full-duplex single-core bidirectional transmission experiment was conducted for 400 GPON, and the effectiveness of shared protection for a half-split WDM-PON was confirmed. An experimental study on the main link failure related to the proposed bypass/backup-link switchable WDM-PON was published in [21], but an experimental study on route switching and wavelength conversion for sub-link failures has not yet been reported. The results of these experiments represent new developments from the previous results.
2. Shared/dedicated switchable P-MP NW with full connection/half split
Figures 1, 2, and 3 show three types of configurations for coherent WDM-PONs with redundant optical circuits before and after a link failure. Table 1 gives the gain of the uplink/downlink transmissions for the three configurations, the number of links in the configuration, the required number of optical switches, and the transmission capacity per link. The optical switch has two specifications: the switching time must be sufficiently shorter than the time required for a person to turn on the switch when there is a link failure [16], and the device insertion loss is less, so a micro electromechanical systems (MEMS) optical switch is assumed. We compare the three configurations: a fully coupled PON with shared protection, a half-split PON with shared protection, and a fully coupled / half-split switchable PON with shared protection that can be reconnected by switching to a secondary link when the primary link fails. When a failure occurs, four points related to disaster resistance are compared that concern the required number of optical devices related to hardware cost, the amount of reduction in branch loss related to power gain, the allowable number of failures of the main link, and that for the spare links. A PON that can be fully coupled/half-divided requires one more 1×2 optical switch than a half-divided PON, but the number of optical fibers in the aggregation section can be reduced. Furthermore, compared to the branch loss of the main/secondary link in fully coupled PON, the branch loss of the main link can be improved by 6 dB and the insertion loss of the sub-link can be improved by 3 dB. Here, the insertion loss of the optical switch is assumed to be 0.5 dB or less, and is not included in this calculation. An inexpensive optical CPL is placed in the PON configuration that can be fully coupled/half-split, and an optical switch group is placed on the OLT side. Due to this asymmetric structure, on the ONU side, coupling and branching are performed similar to that in a conventional PON. On the OLT side, by selecting a route that utilizes the switching of the optical switches, it is possible to switch instantly the connection immediately after a link failure occurs.
Fig. 3. Fully coupled / half-split switchable PON configuration, wavelength, and route allocation (a) uplink, (b) downlink.
Table 1. Comparison of Three Types of NW Configurations
Some laser diodes (LDs) used on the OLT side go into sleep mode and perform power-saving operations when no link failure occurs. When the link fails, the configuration switches to use all LDs, and requires a wavelength conversion function. By sharing the coupling/branching of the optical CPL and the path control function of the optical switch, the group of LDs on the OLT side can reduce the cost during system operation. Figure 4 shows the optical fiber wiring when LDs are shared on the OLT side. There is a bidirectional transmission between the two PONs and no link failure. Figure 5 shows a typical LD configuration on the OLT side. By configuring the switches as shown in Fig. 4(a), two of the four LDs can be put into sleep mode for power savings. Figures 4(b,c) show the optical switch configuration for uplink/downlink transmission of fully coupled/half-divided PONs when a main link and sub-link failure occurs, and all LDs are in operation between the two PONs.
Fig. 4. LD common configuration on the OLT side (a) no link failure case and (b) main link failure case (c) sub-link failure case.
3. Optical switching and wavelength allocation before/after link failure
When using a sub-link during a primary link failure, wavelength conversion is required at the PON on the primary link failure side to avoid wavelength collision between the two PONs. The PONs can be reconnected instantly without suffering from fiber end-face reflection by sharing wavelength allocation information between the two PONs in advance before a link failure, and by immediately switching the route using the optical switch when a failure is detected.
On the ONU side, as shown in Fig. 5, wavelength conversion can be performed while shortening the optical frequency detuning time without wavelength collision by assigning wavelengths before and after a main link failure and immediately after the sub-link failure. When the primary link fails, ONUs A-1 and B-2 are assigned the same wavelength before and after the link failure occurs, and the same wavelength is assigned to the transmitting and receiving sides, so wavelength conversion is not necessary. Therefore, the primary time that must be considered before reconnecting the PONs after a link failure is the path switching time of the optical switch and the control time for the upper layer.
4. Experimental configuration
Figure 6 shows the structure of the experimental system. We conduct a principle confirmation experiment in the presence or absence of link failure for the three types of coherent WDM-PON systems equipped with protection functions. In this confirmation experiment, we emulate the case where no link failure occurs, and the case where one link failure occurs and reconnection is established. We evaluate the static characteristics in each case. In the experiment, the pattern of the wavelength, before and after wavelength conversion, was selected by the WSS to generate the desired WDM signal. A 112 Gbit/s, double-polarized quadrature phase-shift keying signal with a signal format of 50 GHz or 100 GHz and a non-zero return format is used as the WDM signal for the uplink/downlink. The channel spacing and modulated signal conditions in this system are selected assuming the use of a C Form-factor Pluggable (CFP) [22], which is a transceiver compliant with the 100GBASE-ZR standard established by IEEE802.3ct. In this emulation, the signal quality levels of the uplink and downlink transmissions are evaluated when full-duplex transmission of a single core fiber is performed on the main link and sub-link, and wavelength conversion and route switching are performed before and after link failure. The switching time of the MEMS optical switch (Polatis series 6000) used in the experiment is 25 ms. In the optical fiber link failure, end-face reflection is simulated by physically opening up the connecting end faces of the two patch cords and exposing them to air. In addition, we simulated a decrease in received power that changes according to the transmission distance by adjusting the input power to the trunk span using an optical attenuator. At that time, we evaluate the bit error rate (BER) characteristics when only the regular signal power transmitted through the optical fiber is changed while keeping the amount of reflected light due to the failure of the optical fiber link constant. We confirmed in advance that the amount of reflection from the fiber end face when the optical fiber is cut is −15 dB.
Fig. 6. Experimental configuration (a) fully coupled PON, (b) half-split PON, and (c) switchable PON.
In this principle experiment, the trunk span near the ONU side is disconnected to maximize the effect of reflection on the downlink signal. All uplink and downlink signal are regarded as an independent signal generated from different transmitters during full-duplex transmission with the wavelength selection switch selecting the wavelength of the WDM signal and combining the two outputs of the wavelength selective switch (WSS) with a CPL after eliminating the data correlation using fiber lengths with different optical path lengths.
As shown in Fig. 6, the signal in the LD group is modulated by the Dual Polarization - In-phase Quadrature modulator (DP-IQM), and the WSS selects the wavelength. The generated WDM signal is input to the trunk span fiber corresponding to the main link or the sub-link as the uplink transmission or downlink transmission, respectively.
On the Rx side of each ONU and OLT, the signal power is combined with the light of the local oscillator, and only the target wavelength component is extracted from the WDM signal based on coherent reception using homodyne detection. After the coherent receiver separates the data into four lanes, i.e., XI, XQ, YI, and YQ, an analog-to-digital converter extracts sample data asynchronously. The digital signal processor performs polarization separation, frequency offset compensation, and carrier phase offset compensation. After hard judgment processing of the reproduced signal, the BER of the restored bitstream is measured.
5. Experimental results
Based on the experimental model described in the previous section, in a NW configuration with three types of redundant fibers, the BER measurement results of four ONUs in a single core bidirectional connection in which the uplink and downlink are connected simultaneously in the cases of a link connection and failure are shown. In this evaluation, the threshold value of the error correction limit is BER = 3.8 × 10−3, which compensates for error-free operation when a rigid judgment-oriented forward error correction code with 7% redundancy is used. In a NW configuration with three redundant routes, we measure the power penalty when the primary link fails in the downlink transmission and the power penalty when the secondary link fails.
Figures 7 and 8 show the BER characteristics of a fully coupled PON and a half-split PON. When the primary link is broken, the upstream signal is reflected by the end face between the air and the glass in the downlink, which is affected by interference, resulting in a power penalty of 7 dB compared to that before the link failed. In the protection shown in Figs. 1 and 2, under the condition that the number of wavelengths is the same as the number of ONUs, the downlink wavelength and the uplink wavelength match when a link failure occurs, and a 7 dB penalty is incurred because reflection cannot be avoided. In the uplink, the downlink signal is not affected by the interference because the optical path is switched after the link failure. Figures 9 and 10 show the BER characteristics when the primary link fails and when the sub-link fails in a PON that can be switched between fully connected and half-split modes. We confirm that in the cases of main link failure and sub-link failure, it is possible to keep the power penalty within 2 dB compared to that before link failure by appropriately allocating wavelengths according to the route switching. As shown in Fig. 3, even if the number of wavelengths is the same as the number of ONUs, the downlink and uplink wavelengths are different in each ONU when link failure occurs, and reflection can be avoided. However, because the signal used in the 50 GHz interval WDM used this time is in the 28 GSymbol/s NRZ format, a 2 dB penalty is incurred because the reflected signal interferes with the primary signal.
Figure 11 shows the optical spectrum of the reflected signal of the primary signal at the ONU point where the downlink signal is received under three conditions: no-link failure, main link failure, and sub-link failure in a fully coupled/half-split PON. After the link fails, Figs. 11(b,c) corresponds to the optical spectrum when the route is switched, and the wavelength conversion is performed. The optical spectra of the primary signal and the reflected signal are acquired when the uplink and downlink are transmitted independently. When there is no link failure, the primary signal power is 20 dB higher than that for the reflected signal, so the effect of interference due to the reflected signal is small, and single-core bidirectional transmission is possible as long as the effect of interference is tolerated. After the primary link and sub-link fail, the primary signal is 8 dB and the reflected signal is 15 dB, so the influence of interference from the reflected signal increases. It is necessary to avoid the influence of interference from wavelength conversion.
Fig. 11. Optical spectrum of the main signal and end-face reflection at FEC limit (a) no link failure, (b) primary link break, and (c) secondary link break.
6. Conclusion
In conclusion, we proposed a fully coupled/half-split NW selection type WDM-PON configuration that has a shared protection function and can switch between shared/dedicated links. A principle experiment was conducted to confirm the effectiveness of the wavelength allocation based on path switching using an optical switch and the colorless function employing tunable LDs during link failure. In the experiment emulating the link failure, the optical switch was switched, and the wavelength was converted after a single link failure occurred in the same wavelength band single-core bidirectional transmission. We clarified that it is possible to reconnect the PONs while keeping the influence of end-face reflection within 2 dB after the link failure.
Funding
National Institute of Advanced Industrial Science and Technology (JPNP20017).
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. J. S. Wey and J. Zhang, “Passive optical networks for 5G transport: technology and standards,” J. Lightwave Technol. 37(12), 2830–2837 (2019). [CrossRef]
2. R. Borkowski, M. Straub, Y. Ou, Y. Lefevre, Z. L. Jelic, W. Lanneer, N. Kaneda, A. Mahadevan, V. Huckstadt, D. V. Veen, V. Houtsma, W. Coomans, R. Bonk, and J. Maes, “FLCS-PON - A 100 Gbit/s flexible passive optical network: Concepts and field trial,” J. Lightwave Technol. 39(16), 5314–5324 (2021). [CrossRef]
3. Y. Zhu, L. Yi, B. Yang, X. Huang, J. S. Wey, Z. Ma, and W. Hu, “Comparative study of cost-effective coherent and direct detection schemes for 100 Gb/s/λ PON,” J. Opt. Commun. Netw. 12(9), D36–D47 (2020). [CrossRef]
4. N. Suzuki, S. Yoshima, H. Miura, and K. Motoshima, “Demonstration of 100-Gb/s/λ-based coherent WDM-PON system using new AGC EDFA based upstream preamplifier and optically superimposed AMCC function,” J. Lightwave Technol. 35(8), 1415–1421 (2017). [CrossRef]
5. T. Kodama, R. Matsumoto, and N. Suzuki, “Demonstration of data-rate and power-budget adaptive 100 Gb/s/λ-based coherent PON downlink transmission,” Proc. Optical Fiber Communication Conference (OFC), Th2A.22, Los Angeles, USA, Mar. 2017.
6. J. Kani, J. Terada, K. Suzuki, and A. Otaka, “Solutions for future mobile fronthaul and access-network convergence,” J. Lightwave Technol. 35(1), 527–534 (2017). [CrossRef]
7. N. Suzuki, H. Miura, K. Matsuda, R. Matsumoto, and K. Motoshima, “100 Gb/s to 1 Tb/s based coherent passive optical network technology,” J. Lightwave Technol. 36(8), 1485–1491 (2018). [CrossRef]
8. IEEE P802.3ca 100G-EPON Task Force “Physical layer specifications and management parameters for 25 Gb/s and 50 Gb/s passive optical networks” [Online].
9. V. Houtsma and D. V. Veen, “Optical strategies for economical next generation 50 and 100G PON,” Proc. Optical Fiber Communication Conference (OFC), M2B.1, Apr. 2019.
10. P. Chanclou, L. A. Neto, K. Grzybowski, Z. Tayq, F. Saliou, and N. Genay, “Mobile fronthaul architecture and technologies: A RAN equipment assessment,” J. Opt. Commun. Netw. 10(1), A1–A7 (2018). [CrossRef]
11. P. Chanclou, “Optical Fiber Solution for Mobile Fronthaul to Achieve Cloud Radio Access Network,” Proc. Future Network and Mobile Summit, July 2013.
12. T. Pfeiffer, “Next generation mobile fronthaul and midhaul architecture,” J. Opt. Commun. Netw. 7(11), B38–B45 (2015). [CrossRef]
13. ITU-T G Suppl. 51, “Passive optical network protection considerations,” June 2017 [Online].
14. N. Parkin and A. Rafel, “Novel low cost PON protection via harvested power,” Proc. Optical Fiber Communication Conference (OFC), Th2A.35, Mar. 2020.
15. L. B Du, S. Yin, X. Zhao, T. Zhang, A. Barratt, J. Jiang, D. Wang, and C. F Lam, “Redundancy for long-reach TWDM PON,” Proc. Optical Fiber Communication Conference (OFC), Th2A.56, Mar. 2018.
16. S. Zhang, W. Ji, X. Li, K. Huang, and Z. Yan, “Efficient and reliable protection mechanism in long-reach PON,” J. Opt. Commun. Netw. 8(1), 23–32 (2016). [CrossRef]
17. Q. Yang, N. Deng, X. Zhou, and C-K. Chan, “A mobile fronthaul system architecture for dynamic provisioning and protection,” Proc. 21st OptoElectronics and Communications Conference (OECC) held jointly with 2016 International Conference on Photonics in Switching (PS), pp. 1–3, 2016.
18. T. Nishitani, Y. Hirano, M. Noda, and K. Motoshima, “Protection systems for optical access networks,” J. Lightwave Technol. 35(6), 1197–1203 (2017). [CrossRef]
19. T. Kodama, Y. Azuma, and T. Ishikawa, “Colorless and directionless coherent WDM-PON architecture with extended star topology using a self-healing for bidirectional link protection,” Proc. OptoElectronics and Communications Conference (OECC), T1-3.6, Taipei, Taiwan, Nov. 2020.
20. S. Kaneko, T. Yoshida, S. Kimura, and N. Yoshimoto, “Reliable λ-tuning OLT-protection method based on backup-wavelength pre-assignment and discovery process for resilient WDM/TDM-PONs,” J. Lightwave Technol. 33(8), 1617–1622 (2015). [CrossRef]
21. T. Kodama, T. Goto, and R. Matsumoto, “Wavelength collision-free and low-loss full-duplex transmission over switchable full-coupling or half-split coherent WDM-PON system with shared protection,” Proc. Optical Fiber Communication Conference (OFC), Th5I.3, Mar. 2021.
22. Y. Loussouarn, E. Pincemin, M. Pan, G. Miller, A. Gibbemeyer, and B. Mikkelsen, “Multi-rate multi-format CFP/CFP2 digital coherent interfaces for data center interconnects, metro, and long-haul optical communications,” J. Lightwave Technol. 37(2), 538–547 (2019). [CrossRef]
