Open Access
Add to BibSonomy
Abstract
The demand for fifth generation (5G) mobile systems has been exploding. There are strict latency and transmission capacity requirements, especially in the optical fiber section called the mobile fronthaul. A wavelength division multiplexing passive optical network (WDM-PON) system is an attractive solution to accommodate the high demand of traffic. However, wavelength management, which is important in a WDM-PON system, has not been sufficiently studied at the system level. In previous studies, control by a management control channel independent of the client signal, called the auxiliary management and control channel (AMCC), has been considered in many cases. In most of these studies, physical layer evaluation, such as the effect of superimposing the management control channel, was conducted for feasibility studies, and it was only confirmed that the effect was minor by measuring the bit error rate. In this paper, we review and compare the superimposition schemes of the AMCC to clarify operating conditions such as modulation format and modulation index of the AMCC signals. In addition, we propose a WDM-PON system managed using AMCC for ONU activation, wavelength adjustment and alive monitoring. We then discuss the connection of a WDM-PON prototype system we developed to a 5G mobile system to evaluate the effect of throughput on user equipment. The results indicate that the WDM-PON prototype system did not degrade (less than 2%) the downlink throughput even when the wavelength adjustment function was used against wavelength drift.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Fifth generation (5G) mobile systems have launched in limited areas, and beyond-5G is necessary to meet traffic expansion [1]. Mobile telecom operators have introduced the centralized radio access network (C-RAN) concept to address this necessity. This concept involves two types of equipment: a centralized unit/distributed unit (CU/DU) and remote unit (RU). These units are connected via an optical fiber link called a mobile fronthaul (MFH). One of the interfaces for an MFH, the common public radio interface (CPRI), has been defined [2]. In the 4G mobile system, an MFH typically uses the CPRI between CUs/DUs and RUs. However, its optical bandwidth in the MFH reaches more than 10 times that of the original wireless data rate to convert the wireless in-phase and quadrature-phase (IQ) analog signals to digitized signals. If it is assumed that the bandwidth of 5G mobile is 100 MHz and the number of antenna ports is 32, the line rate of the MFH can be 157.3 Gbps for uplink (UL) and downlink (DL) [3]. To loosen the bandwidth requirement, functional split points have been proposed such as a lower-layer split (LLS) or high-layer split (HLS): the RUs execute baseband processing partially or wholly [4], as defined with the enhanced CPRI (eCPRI) [5] or 3GPP [6]. In addition, the required number of optical fibers will increase in future MFH systems. This is because the coverage areas of RUs using millimeter waves are small; therefore, a large number of RUs is required to cover a wide area. Moreover, by taking the requirement for beyond-5G and 6G, it is necessary to reduce the latency to 1 ms in an end-to-end connection [1].
To address these issues, wavelength division multiplexing passive optical networks (WDM-PONs) are an attractive candidate [7–9]. WDM-PONs can reduce the number of optical fibers by sharing the use of feeder fibers among optical network units (ONUs). They can also provide lower latency and higher capacity than a time division multiplexing (TDM)-PON. This is because an optical line terminal (OLT) has a logically point-to-point connection with each ONU in the WDM-PON system. In a TDM-PON system, however, ONUs share a wavelength for each direction: UL and DL. Then, each ONU has to wait for permission from the OLT to send an UL signal [10].
To implement a WDM-PON system for accommodating densely deployed RUs, a colorless ONU, which has a tunable transceiver, is an attractive solution for easy inventory. To install a tunable system in a WDM-PON, controlling the wavelength resources is very important for achieving stable communication to support mobile systems.
To control and manage wavelength resources, the ITU-T Recommendation G.989 series [11–13] specifies an embedded communication channel called the auxiliary management and control channel (AMCC) to manage and control the wavelength. The feasibility of the AMCC has been confirmed and evaluated [14–19]. Our previous studies showed the feasibility of superimposing AMCC signals with negligible effect on client signals such as CPRI signals (options 3 and 7) and 10-Gbps pseudo random bit sequence (PRBS) 231-1 with back-to-back configurations [14,15]. Tayq et al. also confirmed that the effect on the CPRI signals by embedding the AMCC signals was negligible with 20-km transmission [16]. Tan et al. demonstrated 10-Mbps AMCC signal transmission without degrading the client signal by reconstructing the client signal at the receiver side [8]. Yoshima et al. showed that the AMCC was superimposed to 100 Gbps [18]. We also investigated the management of WDM-PON systems with the AMCC such as upstream wavelength adjustment, activation, and protection [20–22].
However, these studies did not discuss the system level confirmation for accommodating 5G mobile systems. The throughput measurement of clien signals superimposing AMCC signals and long-term stability test are required to confirm the system level feasibility. The superimposition of the AMCC signals may affect the exchange of the control signals in a 5G system. If the control signals of a 5G system are affected, the throughput of mobile signals might degrade.
In this paper, we propose a WDM-PON system managed using the AMCC and discuss the methods for establishing the AMCC. We experimentally demonstrated that the WDM-PON prototype system we developed controls the wavelength while not affecting 5G cellular systems. The difference from our previous studies is the trial involving our WDM-PON prototype system for accommodating a 5G mobile system. The remainder of the paper is organized as follows. In Section II, we explain the architecture of the proposed WDM-PON system managed using the AMCC. We also review the methods for establishing the AMCC. In Section III, we present the experiments and trial we conducted to evaluate our prototype WDM-PON system when connected to a 5G cellular system. Finally, we conclude the paper in Section IV.
2. Architecture of proposed WDM-PON system managed with auxiliary embedded management and control channel
2.1 WDM-PON architecture
In this section, we describe our proposed WDM-PON system managed using the AMCC for an MFH, as shown in Fig. 1. The CUs/DUs are connected to OLTs, and RUs are connected to ONUs. We assume that 4G and 5G cellar systems are accommodated in a WDM-PON system for the cost-effective shared use of the deployed optical fiber. OLTs and ONUs are connected via an optical distribution network with an optical power splitter at the ONU side. We assume that an arrayed waveguide grating (AWG) is used as a wavelength multiplexing/demultiplexing filter at the OLT side.
The key point of this system is the flexible management function provided via the AMCC. This is because the client signals (i.e., mobile signals) do not convey the management and control information about the WDM-PON system. The channel requires transparency to accommodate many small cells and macro cells independently from any protocols the cells may use. Because it is desirable to accommodate various types of C-RAN architectures, the new architectures discussed in a previous study [4] are taken into consideration. The OLT has an AMCC controller, and each ONU has an AMCC processing unit. The OLT and ONUs send AMCC messages embedded into the client signals of the DL and UL.
2.2 Implementation of auxiliary embedded management and control channel
2.2.1 Channel methods: in-band channel/out-of-band channel
There are mainly two methods of transmitting management and control information: through the in-band channel and through the out-of-band channel, as shown in Fig. 2. We define the in-band channel as the channel changing the bit stream with the management information. The in-band channel embeds the information into client signals as the overhead. Other ways are defined as the out-of-band channel.
There are two possible solutions to establishing the in-band channel: 1) using the reserved area of the main signal and 2) encapsulating the information of the client signal. An example of using the reserved area of the main signal is the CPRI signals. When the information is input into CPRI signals, the information about the WDM-PON system can be embedded as “vendor specific” or “reserved” [2]. For example, the line rate can be expected to be approximately 120 kbps calculated as follows. 256-bytes control words are sent every 66.76 ms as a hyper frame, regardless of the rate of the CPRI signal. By using 8 bits of the reserved area, the line rate of AMCC signals reaches 120 kbps. This line rate is close to 128 kbps described in the ITU-T recommendation G.989.2. The benefit of this method is that no additional circuit is required to generate management and control signals. The second solution is encapsulating client signals over the Ethernet [23] or over the optical transport network (OTN) [24]. The advantage of this method is that equipment cost can be reduced because the operators can reuse the existing packet switched networks in local area and metro/core networks. However, we need to carefully consider what information should be put into each client signals that the mobile systems use as well as where it should be input. The number of items for management information is limited because each frame structure is defined in the documentation. Moreover, the in-band method requires modifications of the mobile communication equipment that generates the CPRI signal to embed the management and control signal of WDM-PON to the CPRI frame.
The out-of-band channel conveys the management and control information through a different physical path. This method enables the AMCC to be established independently from the frame structure of the client signal. However, we have to add an additional circuit, such as a low pass filter (LPF) circuit, to generate management and control signals and separate them from client signals. To accommodate future mobile networks, the independence of the out-of-band channel is quite promising because future mobile networks will probably use a new C-RAN architecture that may have no compatibility with current C-RAN architectures such as CPRI.
2.2.2 Transmission parameter of AMCC signals
To send AMCC signals in out-of-band channel, the operator has to decide the following parameters: superimposition method, modulation format, line rate of the AMCC, and modulation index. We discuss each parameter below.
There are two methods of superimposing the independent management and control channels on to the client signals. Figure 3 shows a transmission scheme for direct laser modulation (DML). The first method is electrical domain superimposition. The AMCC signals are superimposed to the client signal with an electrical power combiner. A laser diode (LD)-bias pin is also available to superimpose the AMCC signals by changing the bias point of DML. The other method is optical domain superimposition. An external modulator, such as a variable optical attenuator (VOA) or semiconductor optical amplifier (SOA), superimposes the AMCC signals to client signals.
Two modulation formats have been discussed in the ITU-T G.989.2 recommendation as AMCC [12] regarding AMCC signals. The first is baseband over-modulation. At the transmitter side, the management data are superimposed on the client signals as baseband signals. At the receiver side, the received signals are separated into client signals and management signals after low pass filtering, then the desired signals are detected. The second modulation format is the low-frequency pilot tone. At the transmitter side, the baseband signals of the management signals are up-converted to low-frequency signals. Then, there are three choices to modulate the pilot tone: amplitude shift keying (ASK), phase shift keying (PSK), and frequency shift keying (FSK). From a qualitative point of view, when comparing PSK and ASK, it is obvious that binary-PSK (BPSK) has better characteristics when considering the distance in IQ symbol space. In a previous study [25], BPSK showed the most negligible effect on the client signal. On the other hand, ASK can be implemented simply with envelop detection. Thus, the appropriate modulation format should be selected in accordance with the system requirement.
The line rate of AMCC signals must be determined. The candidate line rate of AMCC signals is in the range of kbps because the purpose of the channel is to control and manage the ONUs of a WDM-PON system. A WDM-PON system does not require quick management as with the UL control of a TDM-PON system. Thus, the serial communication system can be a reference such as 100 kbps in Inter-Integrated Circuit (I2C) and 115 kbps in Recommended Standard-232C (RS-232C).
Finally, the modulation index of AMCC signals affects the performance of the client and AMCC signals. A higher modulation index improves the performance of the AMCC signals but degrades that of the client signals, and vice versa. Since this performance depends on other parameters including the coding scheme of the client signal, as shown later, the modulation index should be determined through experimental evaluation.
The modulation format depends on which superimposition method is selected. Baseband over-modulation is only available for optical domain superimposition. This is because the transmitter has a DC-block whose cut-off frequency is typically 10–50 kHz, then the baseband signals in kbps cannot pass through. The low-frequency pilot tone can be superimposed in both domains, but a typical VOA has a low-pass characteristic with a cut-off frequency around 100 kHz.
To cost-effectively control a WDM-PON system, the low-frequency pilot tone superimposed in the electrical domain is promising. The reason is that electrical domain superimposition makes it simple to send AMCC signals.
2.2.3 Receptions of AMCC signals
There are two possible methods of receiving AMCC signals, as shown in Fig. 3. The first method involves distributing the signal in the electrical domain using a divider. The second method involves using a power splitter to separate the signals in the optical domain, which requires two photodiodes (PDs). Thus, extraction at the electrical domain is more attractive because the receiver can detect AMCC signals with one PD, whereas extraction at the optical domain requires two PDs. Due to the optical power splitter for detecting AMCC signals, more loss budget is required for the client signals.
To extract the AMCC signals at the electrical domain, two components might affect the signals. The AMCC signals can pass through the limiting amplifier but cannot pass through the clock data recovery (CDR). Figure 4(a) shows the spectrum of a 500 kHz pilot-tone signal passing through the limiting amplifier. With only the limiting amplifier, the tone signal is present, but with the addition of the CDR, the tone signal is not detected, as shown in Fig. 4(b). This can be because the limiting amplifier does not perform excessive signal shaping in the time axis direction, but in the case of using a CDR, the signal is completely regenerated with such as delay flip flop.
Fig. 4. Spectrum of 500 kHz tone signal at receiver (Rx) (a) with limiting amplifier and (b) with limiting amplifier and CDR.
2.2.4 Coding scheme of client signal
Since the AMCC signals are modulated outside the frequency domain of the client signal, it is important to discuss the coding scheme that relates to the frequency component of the client signal. In this section, we quantitatively describe the pilot tones and client signals are affected by two-line coding schemes (8B/10B line coding and 64B/66B line coding), which are both used in the CPRI.
The 8B/10B line coding scheme is a run-length-limited coding scheme based on specific code groups. The maximum consecutive identical digits (CIDs) are 5 bits, and the digital sum variation (the difference in numbers of “0” and “1”) is limited to 6. Therefore, line coding can fortunately reduce the low-frequencies level. Figure 5(a) shows the spectrum of CPRI option 7 with a 500 kHz pilot tone. The frequency under 10 MHz of the client signal is low-level and the pilot tone is superimposed at a low frequency that does not overlap the client signal spectrum.
Fig. 5. (a) Spectrum of CPRI option 7 with AMCC signal, (b) Spectrum of PRBS 231-1 with AMCC signal.
The 64B/66B line coding scheme consists of 2-bit synchronization headers and scrambled data payload. Thus, 64B/66B line coding requires less overhead than 8B/10B line coding. The length of a CID is not bounded due to a scramble with a polynomial. This leads to higher low-frequency levels than that of 8B/10B line coding. As shown in Fig. 5(b), the frequency components extend into the low-frequency domain (< 10 MHz). The PRBS 231 - 1 is one of the transmit test patterns for 64B/66B line coding, as described in IEEE 802.26. Consequently, greater interference may occur between the client signals and pilot tone signals than when 8B/10B line coding is used. If these characteristics are taken into account, it is necessary to superimpose a pilot tone under 10 MHz.
3. Experimental setups and results
As we discussed in Section II, electrical domain superimposing is promising for transmission of AMCC signals. In this section, we describe two transmission experiments and a trial of our WDM-PON prototype system accommodating a 5G mobile system. The transmission experiments were conducted to confirm that electrical pilot tone can be superimposed onto the client signal with negligible interference. As described in Section II, the modulation index is an important parameter for the signal quality of AMCC and main signal, so it was set as the parameter. In the trial, we verified the proposed wavelength control method via AMCC signals with the 5G mobile system. In these experiments and the trial, it is firstly shown that our WDM-PON prototype system controlled the wavelength with no effect on the 5G cellular system.
3.1 Experimental evaluation of effect of superimposing AMCC signals on client signals
3.1.1 AMCC signals with 8B/10B encoded client signal
Figure 6 shows the experimental setup we used to investigate the effect on the client signals encoded using 8B/10B coding scheme by superimposing electrical pilot tone. At the transmitter side, after long term evolution (LTE) signals are generated by the vector signal generator (VSG), the radio frequency (RF) signals are converted to CPRI signals by an RF to CPRI converter (RCC). The VSG and RCC emulate a part of the CU/DU functions to generate LTE signals. The LTE signal bandwidth is 20 MHz; it is standardized as E-TM 3.1 in 3GPP [26]. We tested the CPRI option 7, the line rate of which corresponds to 9.8304 Gb/s. The AMCC signal was generated using an AWG operating at 32 MSamples/sec (MS/s) with 10-bit resolution. A PRBS 27 - 1 was used as the AMCC bit stream. The AMCC bit rate was set to 128 kb/s on the basis of the discussion in G.989.2 [12]. Its modulation format was ASK to test more stringent conditions than BPSK. The data fed into the AWG were pre-processed offline. Waveform shaping was carried out with a raised cosine filter with a 1.0 roll-off factor. The waveform was up-converted to a carrier frequency to 500 kHz. A 22-MHz electrical LPF was used as an anti-aliasing filter. The CPRI signals and pilot-tone-based AMCC signals were multiplexed with a power combiner. An optical carrier 1552.52 nm from the DFB LD was modulated with a LiNbO3 Mach-Zehnder modulator (LN-MZM) driven by the multiplexed signals. The evaluation of the effects caused by superimposition requires linear operation; therefore, we used an external modulator instead of a DML LD. The driving voltage was set to half the Vπ to achieve linear intensity modulation. We used a modulation index defined as Vpp,pilot/Vpp,client, where Vpp,pilot and Vpp,client are the peak voltages of the pilot tone and client signal, respectively. Each amplitude was measured at the point shown in Fig. 6.
At the receiver side, the transmitted signals were detected using a pin-PD. For CPRI signal evaluation, the detected signals were filtered using a 5th-order Bessel-Thomson filter with 3-dB bandwidth of 8 GHz and fed into a CPRI-to-RF converter (CRC) then into a vector signal analyzer (VSA). To evaluate the pilot tone, the detected signals were first subjected to low pass filtering with a 10 MHz Bessel-Thomson filter, digitized using a digital storage oscilloscope (DSO) at 62.5 MS/s, and finally stored in sets of 1 MSamples. Demodulation and bit error rate (BER) estimation were post-processed offline. The pilot tone was extracted from the main signal by using a digital bandpass filter. Finally, down-conversion and timing recovery were carried out. The BER of the AMCC signal was estimated from a Q factor and calculated from measured waveform amplitude distribution.
We evaluated the effects of pilot tone insertion on LTE signals. The modulation index was varied from 10 to 40% with increments of 10%. The carrier frequency was set to 500 kHz. Figure 7 shows the error vector magnitude (EVM) performance of LTE signals using CPRI option 7. From the transmission results for the different modulation indices in Fig. 7, the performance without the pilot tone (denoted as “w/o AMCC”) is also shown in the graph for comparison purposes. As these results indicate, the effects of pilot tone insertion depend on the modulation index; they increase along with the modulation index. We consider that the interference between them is increased because the increase in the modulation index causes a decrease in the signal-to-noise ratio of the CPRI signals. When a network requires optical power penalties on the client signals to not be greater than 0.5 dB at EVM = 1% (red line in Fig. 7), the modulation index can be up to 30%.
Fig. 7. EVM performance of CPRI option 7 (9.8 Gbps, 8B/10B encoding scheme) with and without AMCC signals.
Next, we evaluated the BER characteristics of the AMCC signals. The results indicate that the estimated BER performances of AMCC signals resulted in error-free transmission (< 1E-12) under all conditions plotted in Fig. 7.
3.1.2 AMCC signals with 64B/66B encoded client signals
We investigated the effects of using 64B/66B line coding on the client signals for CPRI option 8 with the experimental setup shown in Fig. 8. The setup was basically the same as that in the previous experiment with 8B/10B encoding client signals. The setup had a loop-back configuration with a commercial transceiver, which was a small form-factor pluggable module (SFP+). The SFP+ module has a limiting amplifier but there is no CDR in it. Its transmitter is composed of 1550-nm distributed feedback (DFB) LDs with an electrical absorption modulator (EAM), and the receiver has a pin-PD. A pulse pattern generator (PPG) generated PRBS 231-1 as a test pattern of the client signal encoded by 64B/66B encoding, as described in Section II. We tested 10.1376 Gbps, which corresponds to the line rate of CPRI option 8. An error detector (ED) was used to measure the BER of client signals at the receiver side after they were filtered with an 8-GHz LPF.
We evaluated the effects of inserting a pilot tone on client signals. Figure 9 shows the BER performance of the client signals with and without the AMCC signals, where the carrier frequency was 500 kHz. It indicates that the interference between multiplexed signals increases with a higher modulation index. The BER requirement of a client signal is 2.34E-5 (red line in Fig. 9) by considering RS(528/514), which is the Reed-Solomon forward error correction used in CPRI option 8 [2]. These results indicate that the modulation index needs to be set under 20% to achieve error free transmission with negligible effect on the client signals (< 0.5 dB optical power penalty).
Fig. 9. BER performance of PRBS 231-1 (10.1 Gbps, test pattern for 64B/66B encoding scheme) with and without pilot tone.
Next, we estimated the performance of the AMCC signals. The modulation index was also set to 10, 20, 25, and 30%. The BERs estimated from Q factors were respectively BER < 1E-2, BER < 1E-8, BER < 1E-10, and BER < 1E-12. The BER of AMCC signals can be sufficient when it is less than 1E-7 because a bit-error occurs less than once a minute due to the line rate of the AMCC signal. Thus, the modulation index under 20% also satisfies the AMCC signal requirement.
3.2 System trial for WDM-prototype with AMCC
The system trial setup is shown in Fig. 10(a). There was a set of a CU/DU and four RUs connected to an OLT and four ONUs, respectively. We used the AMCC evaluation platform described by Nakagawa et al. [24], which is illustrated in Figs. 10(b) and (c). Each OLT and ONU consists of two field-programmable gate array (FPGA) boards, which process the 10-gigabit Ethernet (10GbE) signals and AMCC signals in real time. A tunable SFP+ module was deployed in each OLT and ONU and changed its output wavelength. In each ONU, a tunable filter was set to select the desired DL signals. AMCC signals were generated as 100-kbps BPSK signals up-converted to 500 kHz carrier frequency. The modulation index was set to 10% as specified in ITU-T G.989.2. Since we prioritized the signal quality of the AMCC and client signals, we used the BPSK signal for its better characteristics. The AMCC signals were superimposed to the client signals on the electrical domain by a power combiner. Each AMCC frame format complied with the ITU-T G.989 series. The AMCC FPGA boards send messages to change the wavelength setting of the filter and transceiver via the I2C/RS232C interfaces on the basis of the PC control. For UL wavelengths, ONUs used 1540.56, 1542.14, 1543.73, and 1545.32 nm. For the DL wavelengths, 1556.55, 1558.17, 1559.79, and 1561.42 nm were used. An AWG was used for wavelength multiplexing and demultiplexing at the OLT side. A power splitter divided the optical signals to each ONU.
Fig. 10. (a) Trial setup of WDM-PON system connected to 5G mobile system, (b) OLT and ONU details, (c) photograph of WDM-POM prototype, and (d) photographs of 5G mobile system used for this trial.
In this trial, we tested three fundamental functions of the general WDM-PON system: ONU activation, wavelength adjustment, and alive monitoring. To test the three functions, we implemented them by using AMCC. In the ITU-T G.989 series, there is description of message formats but there is no description of wavelength adjustment sequence in the PtP WDM-PON system. Therefore, we proposed a new sequence for the simple wavelength control. ONU activation is a function to register a new ONU in the OLT and negotiate the wavelengths for DL and UL between them. We used the ONU activation proposed in a previous study [27], which processed multiple ONUs simultaneously by the random waiting time function in the upstream transmission of the registration request. In addition, each ONU started to sweep the wavelength of the tunable filter in a randomized order. These two randomizing functions make it easy to avoid the collision of upstream signals. For wavelength calibration, we used the function developed in our previous study [22], i.e., wavelength adjustment, which is a function to instruct the UL wavelength from OLT to an ONU to prevent wavelength drift caused by LD deterioration over time. For wavelength adjustment, the OLT use the transmission characteristics of the AWG. A wavelength drift degrades power due to the emitted wavelength shifts from the center wavelength of the AWG. With wavelength adjustment, the OLT monitors the received power degradation from the initial value (ΔP). In this trial, the threshold for starting the wavelength calibration algorithm was set at ΔP = -0.5 dB, and a wavelength transition time of 200 ms was set for each wavelength-adjustment instruction. To improve the speed of wavelength adjustment than our previous work [21], which took more than 10 seconds, the monitoring cycle of the received optical power was optimized to 300 ms by considering the time for wavelength transition and AMCC signal processing. We confirmed that the OLT and ONU maintained connection, which is the alive monitoring function. An optical subscriber unit (OSU) and ONU send monitoring signals to each other periodically.
The 5G mobile system consists of a CU/DU and four RUs. Figure 10(d) shows photographs of each piece of equipment. The functional split point of the MFH was set to option 7-1 that the fast Fourier transform was processed at the RU side [28]. We used 10GbE as the interface of the optical link between the CU/DU and RUs. The parameters are listed in Table 1. This trial was conducted in the lobby of the Fujitsu Technology Square office (18 × 36 m) in Shin-Kawasaki, Japan.
Table 1. Parameters of 5G Mobile System
The evaluation of the impact of WDM-PON insertion on the DL throughput of each UE is shown in Fig. 11. We measured the DL throughput because the throughput of the DL signal is affected by the degradation of the UL signal due to the management and control signals in the UL signal. The UL signal degradation might occur by a synchronization error caused by jitter and wander due to the AMCC signal component. The results indicate that the fluctuations in DL throughput are less than 2%, which is within the normal throughput fluctuation range of wireless systems. Thus, it is considered unaffected by the implementation of our WDM-PON system using AMCC. We used the 10GbE frames to transfer the DL data with the functional split point: option 7-1 as mentioned in [28] while the split point of eCPRI is option 7-2. As eCPRI officially supports eCPRI protocol stack over IP/Ethernet [5], close results can be expected when the eCPRI is used as the interface between CU/DU and RU.
Next, the recovery process by wavelength adjustment is shown in Fig. 12(a). We induced wavelength drift to the ONU, which transmitted at a nominal wavelength of 1545.32 nm. From the occurrence of wavelength drift (Step 0 in Fig. 12(a)), we confirmed that the received power was recovered to within the threshold (-0.5 dB) by two steps of adjustment instructions. The effect on the average throughput of the DL signal is shown in Fig. 12(b). We also confirmed continuous operation for more than five hours under the condition of sending and receiving AMCC signals for alive monitoring. From the above results, we succeeded in correcting the wavelength drift and alive monitoring by using AMCC signals without affecting the 5G mobile system.
Fig. 12. (a) ΔP with wavelength adjustment of upstream signal and (b) average DL throughput of before wavelength drift and after wavelength adjustment.
4. Conclusion
We reviewed and compared the superimposition schemes of the AMCC to clarify operating conditions such as modulation format and modulation index of the AMCC signals. The promising method is the out-of-band channel using a pilot-tone signal with electrical domain superimposition onto the client signals. After the review, we discussed two experiments: BER performance test with two different encoding schemes (8B/10B encoding and 64B/66B encoding) and a trial with our developed WDM-PON prototype system accommodating a 5G mobile system. For the application of the AMCC, we proposed and implemented three functions in our prototype: ONU activation, wavelength adjustment, and alive monitoring. The experimental results indicate that the AMCC signals can be superimposed with negligible effect on the client signals. The feasibility of the proposed WDM-PON system was proven in this trial. The WDM-PON system caused no degradation (less than 2%) of DL-signal throughput of the 5G mobile system. In addition, when a wavelength drift occurred, wavelength adjustment was processed and the DL throughput did not degrade from the initial state.
For further investigation, the design for the higher line rate of client signals (25 Gbps or more) is required, but superimposition methods and applications are mainly independent from the line rate. Therefore, the discussion in Section II is available for higher line rate.
Acknowledgments
This technical paper includes a part of results of “The research and development project for realization of the fifth-generation mobile communications system” commissioned by The Ministry of Internal Affairs and Communications, Japan.
We thank Takuya Kanai, Jun-ichi Kani and Tomoaki Yoshida for their useful discussions and comments.
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. NTT DOCOMO, “White paper 5G Evolution and 6G,” IEEE, Jan. 2020. Accessed: Mar. 04, 2021. [Online]. Available: https://www.nttdocomo.co.jp/english/corporate/technology/whitepaper_6 g/.
2. CPRI Specification V7.0, “Common Public Radio Interface (CPRI); Interface Specification,” 2015. [Online]. Available: http://www.cpri.info/downloads/CPRI_v_7_0_2015-10-09.pdf.
3. NTT DOCOMO, “CU-DU split: Refinement for Annex A (Transport network and RAN internal functional split),” Sophia Antipolis, France, 2016.
4. L. M. P. Larsen, A. Checko, and H. L. Christiansen, “A Survey of the Functional Splits Proposed for 5G Mobile Crosshaul Networks,” IEEE Commun. Surv. Tutorials 21(1), 146–172 (2019). [CrossRef]
5. eCPRI Specification V2.0, “Common Public Radio Interface: eCPRI Interface Specification,” 2019. [Online]. Available: http://www.cpri.info/downloads/eCPRI_v_2.0_2019_05_10c.pdf.
6. 3GPP TR 38.801 V14.0.0, “Study on new radio access technology: Radio access architecture and interfaces,” 2017. [Online]. Available: https://portal.3gpp.org/desktopmodules/Specifications/SpecificationDetails.aspx?specificationId=3056.
7. 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]
8. Z. Tan, C. Yang, Z. Xu, L. Chen, X. Huang, H. Guo, Z. Zheng, F. Zhang, and Z. Wang, “Experimental demonstration for over Mbps baseband-over-modulation AMCC implementation in PtP WDM-PON,” 2018Opt. Fiber Commun. Conf. Expo. OFC 2018 - Proc., no. 1, pp. 1–3, 2018.
9. X. Wang, Y. Ji, J. Zhang, L. Bai, and M. Zhang, “Low-Latency Oriented Network Planning for MEC-Enabled WDM-PON Based Fiber-Wireless Access Networks,” IEEE Access 7, 183383–183395 (2019). [CrossRef]
10. H. Uzawa, K. Honda, H. Nakamura, Y. Hirano, K. Nakura, S. Kozaki, A. Okamura, and J. Terada, “First Demonstration of Bandwidth-Allocation Scheme for Network-Slicing-Based TDM-PON toward 5G and IoT Era,” 2019 Opt. Fiber Commun. Conf. Exhib. OFC 2019 - Proc., pp. 2019–2021, 2019.
11. ITU-T. Rec., “G.989.1 40-Gigabit-capable passive optical networks (NG-PON2): General requirements Amendment 1,” 2013. [Online]. Available: https://www.itu.int/rec/T-REC-G.989.1.
12. ITU-T. Rec., “G.989.2 40-Gigabit-capable passive optical networks 2 (NG-PON2): Physical media dependent (PMD) layer specification - Amendment 1,” 2019. [Online]. Available: https://www.itu.int/rec/T-REC-G.989.2.
13. ITU-T. Rec., “G.989.3 40-Gigabit-capable passive optical networks (NG-PON2): Transmission convergence layer specification,” 2015. [Online]. Available: https://www.itu.int/rec/T-REC-G.989.3.
14. K. Honda, T. Kobayashi, T. Shimada, J. Terada, and A. Otaka, “WDM passive optical network managed with embedded pilot tone for mobile fronthaul,” in 2015 European Conference on Optical Communication (ECOC), Sep. 2015, pp. 1–3.
15. K. Honda, T. Kobayashi, S. Nishihara, T. Shimada, J. Terada, and A. Otaka, “Low-frequency Pilot Tone Management for WDM-PON toward Future Mobile Fronthaul employing 64B/66B Line Coding,” in ECOC 2016; 42nd European Conference on Optical Communication, 2016, pp. 1–3.
16. Z. Tayq, L. A. Neto, P. Chanclou, and C. Aupetit-Berthelemot, “Experimental real time amcc implementation for fronthaul in PtP WDM-PON,” Eur. Conf. Opt. Commun. ECOC, pp. 995–997, 2016, [Online]. Available: https://ieeexplore.ieee.org/document/7767791.
17. Z. Xu, C. Yang, Z. Tan, H. Guo, and F. Zhang, “AMCC Superimposition and Extraction With Interference Elimination for 5G Mobile Fronthaul,” IEEE Photonics Technol. Lett. 30(13), 1214–1217 (2018). [CrossRef]
18. S. Yoshima, T. Katsumata, H. Miura, Y. Noguchi, A. Nagasawa, N. Suzuki, and M. Noda, “Experimental investigation of an optically-superimposed AMCC in 100 Gb/s coherent WDM-PON for 5G mobile fronthaul,” Eur. Conf. Opt. Commun. ECOC, no. 1, pp. 986–988, 2016.
19. K. Honda, T. Kobayashi, S. Nishihara, T. Shimada, J. Terada, and A. Otaka, “Experimental Analysis of LTE Signals in WDM-PON Managed by Embedded Pilot Tone,” IEEE Photonics Technol. Lett. 29(5), 431–434 (2017). [CrossRef] .
20. K. Sone, G. Nakagawa, S. Oda, M. Takizawa, Y. Hirose, and T. Hoshida, “Remote Management and Control of WDM-PON System for Fronthaul in Cloud-Radio Access Networks,” in 2017 European Conference on Optical Communication (ECOC), Sep. 2017, vol. 2017-Septe, no. 1, pp. 1–3.
21. K. Honda, H. Nakamura, K. Hara, K. Sone, G. Nakagawa, Y. Hirose, T. Hoshida, J. Terada, and A. Otaka, “Wavelength Adjustment of Upstream Signal using AMCC with Power Monitoring for WDM-PON in 5G Mobile Era,” in Optical Fiber Communication Conference, 2018, p. Tu3L.4.
22. K. Honda, H. Nakamura, K. Hara, K. Sone, G. Nakagawa, Y. Hirose, T. Hoshida, and J. Terada, “Wavelength control method of upstream signals using AMCC in WDM-PON for 5G mobile fronthaul,” Opt. Express vol. 27(19), 26749 (2019). [CrossRef]
23. M. Bahnasy, H. Elbiaze, and C. Truchan, “CPRI over Ethernet: Towards fronthaul/backhaul multiplexing,” in CCNC 2018 - 2018 15th IEEE Annual Consumer Communications and Networking Conference, Mar. 2018, vol. 2018-Janua, pp. 1–7.
24. V. Eramo, M. Listanti, F. G. Lavacca, and P. Iovanna, “Dimensioning of OTN/WDM Rings for the transport of Ethernet/CPRI Flows in 5G Scenario,” in International Conference on Transparent Optical Networks, Sep. 2018, vol. 2018-July.
25. G. Nakagawa, K. Sone, S. Oda, S. Yoshida, Y. Hirose, M. Takizawa, and T. Hoshida, “Development of Evaluation Platform of AMCC Superimposition on CPRI Signal Transmission for Mobile Fronthaul Network,” in 2017 European Conference on Optical Communication (ECOC), Sep. 2017, vol. 2017-Septe, no. 1, pp. 1–3.
26. 3GPP TS 36.141 version 12.6.0 Release 12, “LTE; Evolved Universal Terrestrial Radio Access (E-UTRA); Base Station (BS) conformance testing,” 2015.
27. K. Sone, G. Nakagawa, Y. Hirose, and T. Hoshida, “Demonstration of Simultaneous Multiple ONUs Activation in WDM-PON System for 5G Fronthaul,” in 2018 Optical Fiber Communications Conference and Exposition (OFC), 2018, pp. 1–3.
28. Small Cell Forum, “Small Cell Virtualization Functional Splits and Use Cases, Release 7.0,” 2016.
