1. Introduction

The specification of fifth-generation (5G) mobile networks has already passed the standardization phase, and these networks are being rolled out around the world since 2020. The 5G technology was driven by the commercial operators to accommodate future data capacity requirements for this customer base, as well as complemented by efficient manufacturing demands from industry in the shape of internet of things (IoT) [1]. On the other hand, the demand for the high-speed passive optical networks (PONs) has been increasing due to the spread of broad services [2-6]. Recently, the IEEE 802.3ca Task Force has commenced discussion of the first 100 Gbps-based PON standard in the form of 100G Ethernet PON (100G-EPON) [7]. This PON can support the first stage of the growing bandwidth demands for forthcoming 5G networks.

Figure 1 illustrates the bandwidth demands for forthcoming 5G mobile front-haul (MFH) and mobile back-haul (MBH) networks based on PON. The future MFH/MBH can operates with a data rate of 100 Gbps per small cell and up to around 1 Tbps per macro cell incorporating multiple small cells. Further, the transmission rate per small cell is itself expected to increase exponentially to a maximum MFH link rate of 1 Tbps, depending on future wireless user rates such as beyond 5G (B5G) and sixth-generation (6G) mobile systems [8]. The wavelength-division-multiplexing (WDM) access techniques could provide huge data capacity, extended coverage, long reach connection, and high flexibility [9]. Thus, it is expected that the performance of PON is enhanced strongly when its operation is based on the WDM technique, especially for mobile network applications [10-15]. In this case, the WDM-PON serves each small cell with a single wavelength for downstream (upstream) signal. The WDM-PON performance will be enhanced further when coherent optical detection is used to recover the data at the user end, which this leads to what is known as coherent WDM-PON. The coherent technology is considered a promising candidate for realizing such higher-bandwidth PONs due to its high receiver sensitivity, superior to intensity modulation/direct detection-based PON [10]. Further the coherent PON easily leaves space for inserting additional wavelength to accommodate additional users (optical node units) with the same optical distribution network. Another advantage of coherent PON is that it is the only technique available right now for carrying a signal at 100 Gbps or more per single wavelength.

Figure 1. Bandwidth demand for 5G MFH/MBF-based PON technology [8]

To support the operation of next generation mobile networks (i.e., advanced 5G, B5G, and 6G), the performance of coherent PON needs to be investigated carefully for the case of delivering data rate from 100 Gbps to 1 Tbps for each WDM channel [8]. This issue is addressed in the paper for two values of , 600 Gbps and 1 Tbps. To increase toward 1 Tbps, one can use a hybrid of techniques, including high-order modulation (HOM), high symbol rate, and dual-polarization (DP) multiplexing. Increasing modulation order means increasing the number of bits carried by each symbol while decreasing the distance between symbols in the constellation diagram. This leads to decrease in the transmission distance. Although, increasing the symbol rate will increase for the same value of modulation order. Most of the coherent optical communication systems and optical networks demonstrated in the literature use not exceeding 100 GSps due to the speed limit of electronic-to-optical (and vice versa) conversion. Increasing the symbol rate requires increasing the speed of the digital-to-analog (DAC) (and vice versa ADC) conversion. Note that increasing either the symbol rate or modulation order has advantages and disadvantages. Depending on the system requirements, some combination of them is likely to be used. This issue is addressed on this paper as related to next generation mobile networks.

It is worth to mentioning here that different research groups have started an investigation to design coherent WDM optical communication system carrying data rates of 1 Tbps per wavelength. Most of their results appeared as a short announcement on multimedia websites [16]. Recently, Buchali et al. [17] have reported the implementation of 128 GSps DAC to enable 1.52 Tbps single-carrier transmission over 80 km of single-mode fiber. Nokia has given a talk in OFC 2021 summarizing their progress in beyond 1 Tbps transmission [18]. Even these works are related to point-to-point optical communication systems, the concepts may be adapted or modified to support the goals to be achieved by future coherent WDM-PON for next generation mobile networks.

2. Related Works

In 2017, Suzuki et al. [10] presented the 100 Gbps per λ-based coherent WDM-PON prototype system. A real-time 100 Gbps coherent transceiver with a simplified digital signal processing (DSP) suitable for access spans was highlighted.

In 2018, Suzuki et al. [8] presented design concepts for a coherent WDM-PON operating with a 100 Gbps data rate per channel. Issues related to miniaturization optimization of DSP-based embedded coherent transceivers were given.

In 2018, Matsuda et al. [19] demonstrated hardware-efficient carrier phase recovery and adaptive equalization for 100 Gbps per λ based coherent WDM-PON systems. Downstream transmission with 32 Gbaud (GSps) DP-quadrature phase shift keying (QPSK) signals and offline DSP was demonstrated.

In 2019, Shbair and Nahal [20] reviewed the latest progress of coherent WDM-PON technology operating with 100 Gbps per λ and investigated the system performance when QPSK and DP-QPSK modulation formats are used.

In 2019, Luo et al. [21] proposed a scheme of coherent ultra-dense WDM-PON for symmetrical operations between the uplink and downlink. They experimentally used a real-time field-programmable gate array-based transceivers were used to demonstrate a field trial operation.

In 2020, Segarra et al. [22] reported a low-cost coherent ultra-dense WDM-PON with 6.25 GHz channel spacing. The optical network unit was made by implementing coherent transceivers that have two paired low-cost distributed feedback lasers.

In 2020, Luo et al. [23] proposed a 100 Gbps coherent ultra-dense WDM-PON structure with EDFA that was placed with high power budget at the OLT side for optical signals amplification for both the uplink and downlink. A real-time 4 × 25 Gbps network at 12.5 GHz channel spacing over 50 km SSMF was experimentally demonstrated.

In 2021, Zhou et al. [24] experimentally demonstrated DSP-free 8 × 10 Gbps 4-level pulse-amplitude modulation transmission for coherent WDM-PON cost-effective with channel spacing of 20 GHz in the C-band. The polarization-independent coherent receiver was used at the ONU.

It is clear from this survey that most of the work was related to about 100 Gbps per λ data rate and to modulation order 16 or less. Next generation networks are expected to operate with more than 100 Gbps per λ data rate (toward 1 Tbps per λ) and high-order modulation format (64- and 128-QAM) to support the required ultra-high data services. These issues will be addressed in this paper.

3. Design Issues and Configurations for Coherent WDM-PONs Supporting 5G and Beyond Services

This section presents the design issue for coherent WDM-PONs to support 5G and beyond 5G (B5G) services. The target is to achieve 600 Gbps and 1 Tbps data rates per single wavelength through using high-order modulations, dual-polarization (DP) 64- and 128-QAM. The designed configurations are implemented using Optisystem software ver. 15.0.

Figure 2 illustrates the schematic of WDM-PON for distributing mobile services. The mobile front-haul (MFH) networks consist of: (i) an optical line terminal (OLT) incorporating an optical multiplexer to collect the data from the baseband network unit (BBU) located in the base station after optical modulations; (ii) optical network unit (ONU) incorporating optical demultiplexer to serve multi optical receiver/antenna units; and (iii) standard single-mode fiber (SSMF) acting as the transmission link between the OLT and the ONU.

Figure 2. Simplified schematic of the WDM-PON for mobile service

The following requirements are needed to be satisfied for the WDM-PONs designed in this work

(i) The WDM system operates in the C band of the optical communication spectrum. This band is located around the 1550 nm wavelength, where the SSMF has a minimum loss of ≈ 0.2 dB/km. Note that the C band is usually used for optical communication systems, and therefore, one expects that most of the required optical devices for WDM-PONs are available in the market.

(ii) The optical channels frequencies of the WDM system are selected according to the frequency-grid standard issued by the International Union Telecommunication-Telecommunication Standardization Sector (ITU-T). This WDM grid specifies the optical frequencies of the used lasers for different optical frequency spacing of 25, 50, …, 200 GHz.

(iii) No active element is inserted in the transmission fiber link, such as optical amplifiers and optical-to-electrical (and vice versa).

(iv) The ONU supports the service to a macro cell incorporating multiple small cells with each cell is related to one of the used WDM channels wavelengths. Each cell has its own remote radio head (RRH), i.e., antenna, covering 3 or 6 sectors. Since the standard WDM system uses channels, where n is a positive integer, and are adopted in the design (see Figs. 3a and b).

Figure 3. Mobile fronthual for coherent WDM PON. (a) 8-small cell configuration. (b) 16-small cell configuration

(v) The symbol rate of each WDM channel should not exceed 100 GSps since it is limited by the speed of the available electronics. To increase the transmission bit rate per channel (i.e., per wavelength) , one should go to higher-order QAM modulation supported by polarization multiplexing technique (i.e., DP transmission).

(vi) The design should be issued first to support 600 Gbps per λ and then extended to support 1 Tbps per λ. These two values of are recommended by different research groups for future WDM-PONs serving 5G and B5G network services.

Few remarks related to the aforementioned points are given the following

(i) The total bit rate served by the MFH link is computed from

(1)

assuming DP transmission over N_ch channels using M-QAM modulation format. The spectral efficiency (measured in bps per Hz) is computed as

(2)

and it is independent on .

(ii) To ensure zero intersymbol interference (ISI) at the input of the electrical decision circuit used in each channel receiver, the symbol pulse shape should have a raised-cosine filter (RCF) spectrum at the input of this circuit. The filter bandwidth (i.e., the symbol message electrical bandwidth) is given by

(3)

where r is the RCF roll-off factor. The ideal case (i.e., Nyquist filter) corresponds to r = 0, which gives an ideal lowpass filtering spectrum and leads to The other extreme is r = 1, which yields a full RCF spectrum having Note that when the symbols are used to modulate the optical carrier, the generated modulation optical signal has approximately an optical bandwidth due to the generation of upper and lower sidebands.

Consider the case of = 600 Gbps. The corresponding symbol rate equals 50 and 42.9 GSps when 64- and 128-QAM formats are used, respectively. To prevent overlapping between neighboring channels at the optical demultiplexer output, should be less than . In other words, the following condition should be satisfied in the design of the demultiplexer

(4a)

Thus, the maximum value of r that can be used is given by

(4b)

Using lasers with channel spacing = 75 GHz from the WDM grid yields = 0.5 and 0.75 for 64- and 128-QAM formats, respectively. Note that one can choose = 100 GHz for the case of 64-QAM formats signaling to yield = 1 (i.e., full RCF) but the spectral efficiency reduces since it is inversely proportional to (see Eqn. 2).

Now consider the second case of Tbps. The symbol rate equals to 83.3 and 71.4 GSps for 64- and 128-QAM formats. Choosing = 100 GHz yields = 0.2 and 0.4, respectively. One can go to = 125 GHz to relax the demultiplexer design while reducing the spectral efficiency to 80%. The corresponding which in this case is 0.6 and 0.75, respectively.

4. System Parameters Used to Produce the Simulation Results of the Designed WDM-PONs

The results are presented in two scenarios corresponding to two data rates per channel, 600 and 1000 Gbps. Each scenario includes two main parameters, modulation format and number of channels. The simulation is executed using Optisystem software ver. 15.0. The maximum transmission length (i.e., maximum reach) is estimated when the bit error rate (BER) at the received side approaches This BER level corresponds to the BER threshold of 7% hard decision (HD) forward error correcting (FEC) code.

The simulation results are presented for two coherent WDM-PONs operating with 600 and 1000 Gbps data rate per channel (i.e., per wavelength ). Each network is designed to support 8- and 16-antenna cells using either dual-polarization (DP) 64-QAM or 128-QAM signaling format. This yields eight PONs under observation. Each PON is labelled by three indices, namely -PON, where and stand, respectively, for data rate in Gbps carried by a single WDM channel, order of the QAM format, and the number of WDM channels. Therefor takes the value of 600 or 1000, M takes the number 64 or 128, and is either 8 or 16. These PONs are also classified into two categories, A or B depending whether the data rate per channel is 600 or 1000 Gbps, respectively. Accordingly

Class-A PONs

• PON-A1 ≡ (600, 64, 8) PON

• PON-A2 ≡ (600, 128, 8) PON

• PON-A3 ≡ (600, 64, 16) PON

• PON-A4 ≡ (600, 128, 16) PON

Class-B PONs

• PON-B1 ≡ (1000, 64, 8) PON

• PON-B2 ≡ (1000, 128, 8) PON

• PON-B3 ≡ (1000, 64, 16) PON

• PON-B4 ≡ (1000, 128, 16) PON

For comparison purposes, results related to a single-channel counterparts (i.e., N_ch = 1) are also presented for both data rates and both signal formats.

The frequencies of the WDM channels are selected in the C band (i.e., 1550nm region) according to International Telecommunication Union-Telecommunicate Standardization Sector (ITU-T) [25]. The ITU offers a frequency grid corresponding to frequency channel spacing of 25, 50, 75, ...GHz. The channel frequency matches the frequency of the unmodulated laser frequency used for this channel. The channel spacing should be selected to be equal to or less than the bandwidth of the modulated optical carrier

(5)

where is the symbol rate, is the bit rate, M is the QAM order, and r is the roll-factor of the raised-cosine filter used for pulse shaping required to achieve zero intersymbol interference (ISI) at the input of the receiver decision circuit. For = 600 Gbps, = 50 and 42.86 GSps for 64- and 128-QAM signaling, respectively. Choosing = 75 GHz for this data rate requires that should be kept less than 0.5 and 0.75 for these formats, respectively. For 1 Tbps data rate networks, = 125 GHz is used for both M = 6 and 7 formatting where = 83.3 and 71.4 GSps, respectively. In this case should be chosen to be less than 0.5 and 0.75, respectively.

From the previous discussion, = 0.2 is chosen in this work for the designed coherent WDM-PONs, which ensures that the bandwidth of the modulated optical carrier is within the corresponding and with enough frequency guard, , between the neighboring channels. For = 600 Gbps and = 60 and 51.4 GHz assuming 64- and 128-QAM, respectively. This design offers a frequency guard of 15 and 23.6 GHz, respectively. These values are to be compared with and 39.3 GHz for = 1 Tbps and = 0.2, respectively. Note, that in this case, = 100 GHz and 85.7 GHz for 64- and 128-QAM signaling, respectively.

It is interesting to introduce a normalized frequency guard as a parameter to assess the overlapping degree between neighboring channels. For 600 Gbps and 1Tbps networks, = 0.2 and 0.32 for 64- and 128-QAM signaling, respectively. Note that

(6)

In the WDM system, the number of multiplexed channels N_ch is chosen as even number with a central channel is corresponding to channel. If the channel indices are labeled from 1 to , channel 5 and 9 correspond to the central channels for 8- and 16-channel WDM systems, respectively. The central channel frequency is set to 193.1 THz (i.e., = 1.554 nm) in the simulation.

Table 1 lists the used channel frequencies for the 16-channel WDM system with = 75 and 125 GHz adapted for = 600 Gbps and 1 Tbps, respectively.

Table 1. Channels frequencies for 16-channel WDM system

Unless otherwise stated, the main system parameters used in the simulation values are listed in Table 2.

Table 2. Values of main system parameters used in the simulation

5. Simulation Results of 600 Gbps WDM-PONs

This section presents simulation results related to coherent WDM-PONs operating with 600 Gbps per λ data rate. Results are presented for both 64- and 128-QAM signals and for two numbers of multiplexed channels, = 8 and 16. Results for a single-channel PON are also included for comparison purposes.

5.1. 8-Antenna 600 Gbps PON

This subsection presents simulation results for 600 Gbps per λ PONs that use 8-channels WDM system. Two PONs are considered here depending on whether 64-QAM or 128-QAM signal format is used

• PON-A1 ≡ (600, 64, 8) PON

• PON-A2 ≡ (600, 128, 8) PON

Note that the total data rate transmitted over the transmission link is 8 x 600 Gbps = 4.8 Tbps. A frequency channel spacing and roll-factor used in the simulation for both PONs are 75 GHz of 0.2, respectively. The following remarks are taken in the simulation.

(i) Each demultiplexed channel at the receiver side uses its own DP DSP whose parameters are shared with other DSPs used for other demultiplexed channels (Fiber length, SSMF parameters, ...), but it has its own channel frequency.

(ii) Each channel DSP is enabled to yield GVD compensation for the transmission link estimated at the operating channel frequency. Note that the GVD parameter D is varied slightly with the operating wavelength. All the receivers DSPs use D = 17 ps/(nm.km) and dispersion slope S = 0.075 ps/nm2/km at 1550 nm reference wavelength. These data are used by the DSP to estimate the GVD at the channel wavelength according to Eqn. (7).

(7)

(iii) The BERs of all 8 channels are kept under observation at the receiver side. The maximum reach is estimated when the BERs of all the channels do not exceed the threshold level of 4.6x10^-3.

5.1.1. Simulation Results of (600, 64, 8) PON

PON-A1 uses 8-channels WDM system to transmit 600 Gbps DP 64-QAM signal per channel from the base station to the cell. The network is simulated and the results are shown in Fig. 4. Parts a-d show the spectrum of three transmitted channels (i.e., first channel Ch1, central channel Ch5, and last channel Ch8) along with the WDM multiplexed signal. The spectrum of the received WDM signal is given in part e of the figure, when the fiber length is set to 100 km which corresponds to the maximum reach. The spectrum and constellation diagrams corresponding to three demultiplexed signals (Ch1, Ch5, and Ch8) are illustrated in part (f, g, and h) and (i, j, and k), respectively. The BERs of these three channels are and respectively; all BERs are less than BER_th.

Figure 4. Signals spectra and received constellation diagrams related (600, 64, 8) PON

The variation of BERs of the 8 channels with transmission distance (i.e., length of the SSMF link) is illustrated in Table 3. At L = 100 km, the BERs of all the channels are below BER_th = Increasing L to 105 makes the BER of Ch1 (i.e., BER1) more than BER_th. Therefore, the maximum reach is taken according to this channel. Increasing L further to 110 km makes, another channel (Ch2) not satisfy the required BER level. Note that PON-A1 can support the transmission to 8, 7, 6, and 1 antenna when L=100, 105, 110, and 115 km, respectively. At L= 120 km, the BERs of all the received channels are higher than BER_th.

Table 3. Variation of BERs of the eight channels with transmission distance for (600, 64, 8) PON

Figure 5 illustrates the dependence of the BERs of channels 1, 5, and 8 with the transmission distance. Note that Ch5 supports a longer transmission distance than channels 1 and 8. While channel 8 supports a longer distance compared with Ch1. It is clear that Ch1 determines the maximum reach of this network.

Figure 5. Variation BERs of channels 1, 5, and 8 with a transmission distance for (600, 64, 8) PON

5.1.2. Simulation Results of (600, 128, 8) PON

The simulation is repeated for PON-A2 that uses 600 Gbps DP 128-QAM signal over 8-channels WDM system. The results are depicted in Figs. 6 and 7 and Table 4. The spectra of the signals at different points of the system are given in Fig. 6, along with the receiver constellation diagrams. The transmission length is set to the maximum reach

Table 4. BERs variation of 8 channels with a transmission distance (600, 128, 8) PON

Figure 6. Signals spectra and received constellation diagrams related (600, 128, 8) PON

Figure 7 illustrates the dependence of the BERs of channels 1, 5, and 8 with the transmission distance. Note that channel 5 supports a longer transmission distance than channel 1 and 8. Further, Ch8 supports longer distance compared with Ch1. It is clear that Ch1 determines the maximum reach of this network.

Figure 7. Dependence of BERs of channels 1, 5, and 8 on transmission distance (600, 128, 8) PON

Table 4 lists the BERs of the 8 channels for different values of fiber length L. The results in this table highlight the following facts

(i) The maximum reach for the network is 85 km. Operating with L≤ 85 km ensures that all received channels BERs are less than BER_th.

(ii) Increasing L to 90 km makes channel 1 operate unsatisfactory, giving received BER > BERth.

(iii) Increasing L further does not alters the BER picture effectively till L approaches 100 km. At L= 100 km, two additional channels (Ch2 and Ch8) get out of the required-BER operation (i.e., BER > BERth).

(iv) The network can support the required BER condition for 8, 7, and 5 antennas when L = 85, 105, and 110 km, respectively. At L = 115 km, the BERs of all the received channels are less than BER_th.

5.2. 16-Antenna 600 Gbps PON

This subsection presents simulation results for PON-A3 and PON-A4, which support 600 Gbps per λ data rate and 16-Channel WDM transmission under 64- and 128-QAM signaling, respectively.

5.2.1. Performance of (600, 64, 16) PON

PON-A3 is denoted by code (600, 64, 16) PON and its simulation results are depicted in Figs. 8 and 9 along with Table 5. The results in Fig. 8 are plotted for a transmission link length of 60 km corresponding to the maximum reach. The spectra of the Ch1, Ch9, and Ch16 are presented in this, assuming Ch9 is the central channel. The received BERs are and for Ch1, Ch9, and Ch16, respectively.

Table 5. BERs variation of 16 channels with transmission distance (600, 64, 16) PON

Figure 8. Signals spectra and received constellation diagrams related (600, 64, 16) PON

Figure 9 indicate that Ch1 determines the maximum reach of this network. Further, Ch9 gives the best transmission performance supporting longer transmission distance in compression to the other channels. Whereas, the Ch16 transmission distance is longer than Ch1.

Figure 9. BERs of channels 1, 5, and 8 versus transmission distance (600, 64, 16) PON

Table 5 clarifies the variation of BERs with the transmission distance of the 16 channels. When L ≤ 60 km, the BER is less than BER_th for all the 16 channels. Rising L to 65 km makes Ch1 BER more than the BER_th. Increasing L to 90 km makes the BERs of the first and last three channels above BER_th.

5.2.2. Performance of (600, 128, 16) PON

Simulation results related to PON-A4 are given in Fig. 10, Fig. 11, and Table 6. The maximum reach for this network is 45 km and determined by the behaving of Ch1. The signal spectra of three transmitted channels (Ch1, Ch9, and Ch16) at different points of the system are listed in Fig. 10. Also, the corresponding received constellation diagrams of (x polarization) are included.

Table 6. BERs variation with a transmission distance (600, 128, 16) PON

Figure 10. Signals, spectra and received constellation diagrams related (600, 128, 16) PON

Figure 11. BERs of channels 1, 9, and 16 versus transmission distance (600, 128, 16) PON

Figure 11 shows the effect of transmission distance on BERs of channels 1, 9, and 16. Note that center channel Ch9 has the longest transmission distance compared to all system channels. Recall that Ch1 determines the longest transmission distance support of the system.

Table 6 lists the BERs of the 16 channels as a function of fiber length. The results in this table show that using L = 45 km keeps the BERs of all channels under the BER_th. Increasing L to 50 km makes the BER of Ch1 above the threshold limit (BER1>BER_th). Using L =65, 70, and 75 km makes the first three and the last two channels BER above the BER_th. While channels (Ch4 to Ch14) still satisfy the required .

6. Simulation Results of 1000 Gbps WDM-PONs

In this section, simulation results of coherent WDM-PONs operating with 1000 Gbps per λ data rate are given. In addition, the results include two QAM signals, 64 and 128, and for two multiplexed channels, N_ch = 8 and 16. Moreover, to prove the performance, a single-channel PON is involved.

6.1. 8-Antenna 1 Tbps PON

In this subsection, the results for 1000 Gbps per λ PONs and 8-channels WDM system are presented for the following cases

• PON-B1 ≡ (1000, 64, 8) PON

• PON-B2 ≡ (1000, 128, 8) PON

The total transmission data rate over the link is 8 × 1000 Gbps = 8 Tbps. For both PONs, the frequency channel spacing and roll-factor utilized in the following simulation are 125 GHz and 0.2, respectively. Note that the received BERs of all the WDM channels are kept under observation during the simulation. The maximum reach () is estimated when all the channels BERs do not exceed the threshold level of .

6.1.1. Simulation Results of (1000, 6, 8) PON

PON-B1 utilizes 8-channels WDM system to transmit 1000 Gbps DP 64-QAM signal per channel from the base station to the cell. Figure 12 shows the results of the simulation. Parts a-d of this figure depicts the spectrum of three transmitted channels (i.e., first channel Ch1, central channel Ch5, and last channel Ch8) along with the WDM multiplexed signal. In addition, part e shows the spectrum of received WDM signal when the fiber length is set to 40 km, which corresponds to maximum reach. The spectrum and constellation diagrams corresponding to three demultiplexed signals (Ch1, Ch5, and Ch8) are illustrated in parts (f, g, and h) and (i, j, and k) of the figure, respectively. The BERs of these three channels are and respectively; all BERs are less than BER_th.

Figure 12. Signals spectra and received constellation diagram related (1000, 64, 8) PON

The dependence of the BERs of channels 1, 5, and 8 on the transmission distance are shown in Figure 13. Note that Ch5 supports a longer transmission distance than channel 1 and 8. However, Ch8 supports longer distance compared with Ch1. It is clear that Ch1 determines the maximum reach of this network.

Figure 13. BERs of channels 1, 5, and 8 with transmission distance FOR (1000, 64, 8) PON

Table 7 lists the variation of BERs for the 8 channels with transmission distance (i.e., length of the SSMF link). At L = 40 km, the BERs of all the channels are below . Increasing L to 45 km makes the BER of channel 1 (i.e., BER1) more than BER_th. Therefore, the maximum reach of 40 km is taken for this network. Increasing L further to 55 km makes another channel (Ch2) does not satisfy the required BER level. At L= 70 km, the BERs of channels 3 to 6 are less than the .

Table 7. BERs variation with a transmission distance for (1000, 64, 8) PON

6.1.2. Simulation Results of (1000, 128, 8) PON

The simulation is repeated for PON-B2 that uses 1000 Gbps DP 128-QAM signal over 8-channels WDM system. Figures 14 and 15 and Table 8 depict the obtained results. Figure 14 shows the spectra of the signals at different points of the system, along with the receiver constellation diagrams. The transmission length is set to the maximum reach

Table 8. BERs variation with a transmission distance of (1000, 128, 8) PON

Figure 14. Signals spectra and received constellation diagram related (1000, 128, 8) PON

Figure 15 illustrates the dependence of the BERs of channels 1, 5, and 8 on the transmission distance. Note that Ch5 supports a longer transmission distance than channel 1 and 8. While Ch8 supports a longer distance compared with Ch1. It is clear that Ch1 determines the maximum reach of this network.

Figure 15. Varitation of BERs of channels 1, 5, and 8 with a transmission distance of (1000, 128, 8) PON

Table 8 lists the BERs of the 8 channels for different values of fiber length L. The results in this table highlight the following facts

(i) The maximum reach for the network is 30 km. Operating with L≤ 30 km ensures that all received channels BERs are less than BER_th.

(ii) Increasing L to 35km makes Ch1 to operate improbably, giving received BER > BERth.

(iii) Increasing L further does not alter the BER picture effectively till L approaches 60 km. At L= 60 km, two additional channels (Ch2 and Ch8) lags out of the required-BER operation (i.e., BER > BERth).

(iv) The network can support the required BER condition for 8, 7, and 5 antennas when L= 30, 55, and 60 km, respectively.

6.2. 16-Antenna 1 Tbps PON

This subsection presents the results of simulation for 1000 Gbps per λ data rate and 16-channel WDM transmission assuming 64- and 128-QAM signaling represented by PON-B3 and PON-B4, respectively.

6.2.1. Simulation Results of (1000, 6, 16) PON

Figures 16 and 17 along with Table 9 depict the simulation results for PON-B3, which is denoted by code (1000, 64, 16) PON. Figure 16 shows the results corresponding to transmission link length of 20 km (maximum reach). The spectra of the Ch1, Ch9, and Ch16 are presented in this figure, assuming Ch9 is the central channel. The received BERs are and for Ch1, Ch9, and Ch16, respectively.

Table 9. BERs variation with a transmission distance of (1000, 64, 16) PON

Figure 16. Signals spectra and received constellation diagram related (1000, 64, 16) PON

Figure 17 indicate that Ch1 determines the maximum reach of this network with L = 25 km. Further, Ch9 gives the best transmission performance compared to other channels. On the other side Ch16 support transmission distance more than Ch1.

Figure 17. Variation of BER with a transmission distance of Ch1, Ch9, and Ch16 of (1000, 64, 16) PON

The BER variation with a transmission distance of the 16 channels is illustrated in Table 9. The first state in the table of transmission distance indicates that all BERs of the 16 channels are operating lower than BER_th at L = 20. Increasing L to 25 km makes Ch1 BER higher than BER_th. Increasing L to 30, 35, and 50 km makes only 11, 9, and 6 antennas to operate below BER_th, respectively.

6.2.2. Simulation Results of (1000, 128, 16) PON

Simulation results of PON-B4, which is denoted by code (1000, 128, 16) PON are depicted in Figs. 18 and 19 and Table 10. The corresponding maximum reach of transmission link length is 15 km. The spectra and constellation diagram results are plotted in Fig. 18 for a 15 km transmission distance. The spectra of the Ch1, Ch9, and Ch16 are also presented in this figure, assuming Ch9 is the central channel. The received BERs for three channels Ch1, Ch9, and Ch16 are 4.12 x10^-3, 3.47 x10^-4, and 3.37 x10^-3, respectively.

Table 10. Variation of BERs of the 16 channels with a transmission distance of (1000, 7, 16) PON

Figure 18. Signals spectra and received constellation diagram of (1000, 128, 16) PON

Figure 19 shows the dependence of the BERs of Ch1, Ch9, and Ch16 on the transmission distance. Note that Ch1 sets the maximum reach distance of the system. Whereas, Ch9 represents the central channel and supports the longer transmission distance from all other channels. Further, Ch16 has a longer transmission distance than Ch1.

Figure 19. BER variation with a transmission distance of Ch1, Ch9, and Ch16 of (1000, 128, 16) PON

Table 10 illustrates the variation of BERs for the 16 channels with transmission distance. At L = 15 km, the BERs of all the channels are below . Increasing L to 20 km makes the BER of channel 1 and channel 16 more than BER_th. Increasing L further to 25 km makes other channels (Ch2, Ch3, and Ch15) not satisfy the required BER level. At L= 45 km, the BERs of channels 7 to 11 are only less than the BER_th.

7. Performance Comparison

This section presents a brief performance comparison for the PONs investigated in this chapter. The performance measure used for comparison purpose is the maximum reach corresponding to of . Recall that the PONs have been investigated for two modulation formats (64- and 128-QAM), two numbers of multiplexed WDM channels (600 Gbps and 1 Tbps).

Figures 20a and b show chart bars corresponding to as a function N_ch and modulated formats assuming 600 Gbps and 1 Tbps data rate per channel, respectively.

Figure 20. Variation of maximum reach with the number of channels.

Table 11 is deduced for Figs. 20 a and b and lists the ratio which is denoted here by the parameter The results are given for different values of and . Note that going from 64-QAM to 128-QAM signaling yields . This indicates that the maximum reach reduces to 75%.

Table 11. Effect of modulation format on the performance of the designed PONs

An alternative look for performance comparison as depicted in Table 12. According to this table, one can say 16-cahnnel PON reduces to about 50% of the 8 channels PON.

Table 12. Effect of the number of WDM channels on the performance of the designed PONs

It is worth to extending the comparison measure to include the parameter "Total bit rate-maximum reach product", which is computed as and denoted here by the parameter "BLP". Note that The results corresponding to this parameter is given in Table 13 investigating the results in this table highlights the following findings

(i) Operating with = 600 Gbps offers higher BLP compared with = 1 Tbps assuming the same number of WDM channels and modulation format.

(ii) Higher BLP is achieved when the PON is designed with 64-QAM signaling rather than the 128-QAM signaling.

(iii) For fixed and modulation format, the 16-channel PON offers higher BLP compared with 8-channels PON.

(iv) The highest BLP is 576 Tbps.km, which is achieved in a PON designed with Nch = 16, =600 Gbps, and 64-QAM format.

Table 13. Dependence of total bit rate ratio-maximum reach product (BLP) on modulation format, number of WDM channels and bit rate per channel

8. Conclusions

The transmission performance of coherent WDM-PON has been investigated to support 5G and B5G service distribution. The PON sends (receives) the data to (from) a macro cell consisting of 8 or 16 small cells; each small cell uses one of the WDM wavelength. 64- and 128-QAM signaling have been used to support data rate transmission of 600 Gbps and 1 Tbps per single wavelength. Simulation results based on Optisystem software ver. 15.0 have been obtained to determine the maximum reach for each WDM-PON designed in this work. The main conclusions drawn from this study are

(i) WDM channel spacing of 75 and 125 GHz is suitable to transmit 600 Gbps and 1 Tbps data rate, to each small cell, respectively. These values of are applicable for both DP 64- and 128-QAM signaling.

(ii) Adapting the parameters of each channel receiver DSP makes the DSP capable to compensate the effect of fiber dispersion on the channel performance. This statement is valid for both values of 600 Gbps and 1 Tbps, and for both modulation formats.

(iii) When both and the number of WDM channels Nch are kept fixed, the use of 128-QAM format rather than 64-QAM format reduces the maximum reach to approximately 75%. For example, when = 600 Gbps and Nch = 16, reduces from 60 to 45 km when 128-QAM signaling is used in place of 64-QAM. These values of are to be compared with 20 and 15 km, respectively, when Tbps.

(iv) When both modulation format and are kept constant, the maximum reach reduces by 50% approximately when the WDM-PON is designed to serve 16-small cells configuration rather than 8-small cell configuration. For example, using 64-QAM signaling and = 600 Gbps yields = 100 and 60 km when Nch = 8 and 16, respectively. These values are to be compared with = 85 and 45 km, respectively when 128-QAM signaling is used.

(v) Going from = 600 Gbps to 1 Tbps reduces to approximately the third. Using Nch = 16 and 64-QAM signaling yields = 60 and 20 km when = 600 Gbps and 1 Tbps respectively. If 128-QAM signaling is used, = 45 and 15 km, respectively.

(vi) The DP 64-QAM modulation formats can support 1 Tbps transmission if the symbol rate of 84 GSps is used. If extra is transmitted as a header, one can go to 128-QAM signaling at the same value of . this yields 1/7 header.

(vii) Higher bit rate-length product (BLP) is obtained when one uses = 600 Gbps rather than 1 Tbps, 64-QAM rather the 1280-QAM, and = 16 rather than 8. The highest value of BLP is 576 Tbps.km which is achieved in a PON designed with = 16, = 600 Gbps, and DP 64-QAM format.