This paper experimentally demonstrates a novel dynamic λ-OFDMA architecture with selective differential phase shift keying (DPSK) overlaid and colorless optical network unit (ONU). The wavelength can be dynamically assigned to the splitter node in demand through a group of optical switches. Multicast control is realized through an alternative local oscillator at the optical line terminal (OLT). The 10-Gb/s × 4-channel 16QAM-OFDM unicast signal with 2.5-Gb/s DPSK overlaid and 1.25-Gb/s upstream signal have been transmitted over 25-km fiber successfully.
© 2011 OSA
Next generation optical access networks are expected to carry multiple broadband services to customers which offer advantages including high data bandwidth, enhanced security and flexible scalability [1–7]. Several optical access technologies have been proposed to accommodate the bandwidth demand in multi-user access network, such as time division multiple access (TDMA), wavelength division multiplex (WDM), optical code division multiple access (OCDMA) and optical orthogonal frequency division multiplex (O-OFDM) technology. Among them, O-OFDM technology is an effective solution, and it has gained much attention due to its flexible bandwidth and the resistance to chromatic dispersion [8–10]. At the present time, optical access network with highest data rate is achieved by the optical OFDM technology, where 108-Gb/s OFDM downstream signal over a single wavelength is demonstrated employing the polarization multiplexing and direct detection . As for real-time OFDM access, the highest rate of 11.25 Gb/s with 128QAM mapping is achieved by J.M. Tang et al. in .
There are many researches about OFDM access technologies reported such as WDM-OFDM access and TDM-OFDM access (OFDMA) [13–18]. However, WDM-OFDM access is abundance of capacity but requires multiple transceivers and optical filters, which will greatly increase the system cost. Also, whenever the user shuts down the connection, the corresponding transceiver is idle and cannot serve other users, which waste the network resource. On the other hand, due to the A/D bandwidth limitation and time-sharing transmission of upstream, conventional OFDMA is lack of channel scheduling and enough bandwidth for individual users. Ideal solution is to employ wavelength, electrical subcarrier and time slot allocation simultaneously. A multi-wavelength OFDM access network with dynamic wavelength distribution can provide both network capacity and resources efficiency.
Besides, In order to realize more flexible network, a number of studies have been carried out to simultaneously provide both unicast data and broadcast/multicast services [3–6]. However, these schemes cause excess cost to allocate the multicast services, such as complicated time control, multi radio source or additional light source. An orthogonal scheme has been proposed by our previous work , which uses different dimensionalities of one optical carrier to carry the unicast data and multicast data. Nevertheless, our previous scheme is polarization sensitive and has no selective control for multicast due to the polarization shift keying (Polsk) modulation and fixed local oscillation.
In this paper, we propose a novel dynamic λ-OFDMA architecture which simultaneously supports both unicast and selective multicast overlaid transmission, which can solve the above problem about capacity and channel scheduling. In our proposed scheme, differential phase shift keying (DPSK) signal is used for the downstream multicast signal, which can eliminate the cross-talk between the unicast signal and overlaid signal due to its constant energy per bit. The selective scheme is realized by controlling the frequency of local oscillation (LO) at the optical line terminal (OLT). 10-Gb/s × 4-channel transmission is realized in our experiment to validate the feasibility of the proposed scheme. The 10-Gb/s 16QAM-OFDM unicast signal, 2.5-Gb/s DPSK overlaid and 1.25-Gb/s on-off keying (OOK) upstream signal have been successfully demonstrated in our experiment.
2. System architecture
Figure 1(a) illustrates the conventional OFDMA configuration with DPSK multicast data. A Mach-Zehnder modulator (MZM) is used to generate the up-converted OFDM intensity modulated unicast signals. The OFDM baseband signal is mixed with a LO source to produce an electrical signal that drives the MZM. The central peak of the wavelength has no information added so as to reserve for the upstream data. Then the OFDM signal is sent to a phase modulator (PM) for DPSK overlaid modulation. For each ONU, it is allocated with one transport pipe including different subcarriers and time slots, which is shown as inset in Fig. 1(a). Based on Fig. 1(a), we propose a dynamic λ-OFDMA architecture which can offer enough bandwidth for each ONU and ensure the network flexible and effective resource utilization. The architecture is depicted in Fig. 1(b). There are four λ-channels at the OLT, which can be extended by increasing the number of λ-channel. The total number of ONU is n. In each λ-channel, the system mechanism is the same as Fig. 1(a) shows. The ONU which use the same wavelength share the bandwidth on different subcarriers and time slots of OFDM signal. All the unicast OFDM signals are combined by a MUX and sent to the phase modulator (PM) for DPSK overlaid modulation. Figure 2 shows a sketch map of the downstream and upstream spectra. For the overlay-enable service, the frequency of LO is sent to a higher value (5 GHz), which ensures the OFDM signal spectrum out of the spectral range of DPSK signal. When the phase information partly converted into amplitude information during transmission, the cross-talk between unicast signal and overlaid signal can be significantly eliminated. For the overlay-disable service, the LO of the unicast signals are set to a lower frequency (2.5 GHz), which makes the OFDM spectrum partly overlap on the DPSK one after demodulation, so the DPSK overlaid cannot be recovered correctly at the receiver.
At the optical distribution node (ODN), an arrayed waveguide gratings (AWG), a splitter node and an optical switches (SW) array are used to dynamically allocate the λ-channel to different ONUs. The splitter node is consisting of power splitter groups with different split ratio from 1:8 to 1:64. The 1 × 4 SW after the arrayed waveguide gratings can assign wavelength to any of the splitter groups. The n×n (n input ports and n output ports) SW array is used to allocate the downstream to n ONUs (there we assume n=120). The SW array can be realized on a small chip with the help of micro electro mechanical systems (MEMS) technology [20–22] and it can be controlled through a programmable module. The control information comes from the OLT through a control channel. The schematic of the SW chip is shown as inset in Fig. 1(b), which is consists of a MEMS mirrors, a focus reflector and input/output ports with beam collimation. The MEMS mirrors are tilted as switch elements, which is actuated through the programmable module. The input and output parts are symmetrically folded by the focus reflector. The collimated optical signals from input ports are reflected by the tilted MEMS mirrors of the right side onto the left side one through the focus reflector, and the tilted MEMS mirrors on the left direct the optical signals into the output ports. By adjusting the angle of the beam collimation and mirrors, an input port can connect to any of the output port. For an SW chip with 128 × 128 size, the insertion loss is less than 2.6 dB . A commercial 4-channel arrayed waveguide gratings has a general insertion loss of 1.6 dB, and the additional loss of the splitter with different split ratio is between 0.4 dB and 1.2 dB, so the total insertion loss of the ODN is between 4.6 dB and 5.4 dB without considering the split loss. Otherwise, the insertion loss of ODN for each ONU is between 12.6 dB and 23.4 dB. By monitoring the network traffic flux, the SW array can dynamically change the status of the output ports which connect the ONUs. When the number of users is small, ODN would connect splitter group with low split ratio to ONU, which can increase the optical power for each ONU. Thus the power budget is improved. Besides, for the ONUs requiring low bandwidth, ODN would choose splitter node with high split ratio (1:64 for example); while for the ONUs requiring large bandwidth, ODN can assign them splitter group with lower split ratio(1:16 for example). Hence the resources can be efficiently utilized.
At the ONU side, the downstream signals are sent to two parts through a 3-dB coupler. One part is for the OFDM unicast signal and overlaid signal receivers respectively after passing another coupler. The other part is used for colorless upstream re-modulation  by an intensity modulator (IM). This scheme is easy to implement and can be low power consumption because there is no additional light source at the ONU.
3. Experimental setup and results
The experimental setup is shown in Fig. 3 . In the OLT, the four optical sources are generated by DFB lasers at 1551.20 nm, 1552.20 nm, 1553.22 nm and 1554.21 nm respectively. The 10-Gb/s OFDM unicast signal is produced through MATLAB programming. The time domain OFDM signal is made up of 512 subcarriers with 16QAM mapping, among which 416 subcarriers are used for data, 16 pilot subcarriers are used for phase estimation, and the other is set to zero for over-sampling. The cyclic prefix (CP) length is 1/16 of the time domain. Training sequence is employed in each 40 OFDM frame in order to enable the time synchronization and channel estimation. The electrical waveform is generated through a Tektronix AWG7122B Arbitrary Waveform Generator at 10-bits D/A convertor. Then the OFDM signal is mixed with a 2.5/5-GHz sinusoidal wave by a mixer to execute the up-conversion, where we can realize the selective multicast overlaid through adjusting the frequency of the LO. The measured RF OFDM spectrum is shown in Fig. 4(d) , where the frequencies of the subcarriers are distributed from 3.75 GHz to 6.25 GHz, and the bandwidth of the signal is 2.5 GHz. The MZM is driven by the 10-Gb/s RF OFDM signal with peak-to-peak voltage of 4 V corresponding to half-wave voltage of 5 V. Then the signal is launched into a PM for multicast signal modulation. The signal at 2.5 Gb/s with a PRBS of 231-1 is exclusive-OR (XOR) pre-coded before modulated onto the unicast signal. For the overlay-enable channel, the LO is set to 5 GHz to avoid the overlap of the spectra of the two downstream signals; at the ONU, a low pass filter is employed to filter out the DPSK overlaid. If the end-users cease the subscription for multicast service, then the frequency of the LO would be set at 2.5 GHz, which makes it impossible to recover the multicast signal any more. The downstream optical signals are then boosted to 10 dBm by an EDFA before launching into the 25-km standard single mode fiber (SMF). After transmission, the controllable AWG (cAWG) can allocate the designated wavelength to different output. The free spectrum range (FSR) of the cAWG is 3.2nm and the intrinsic loss at the center frequency is 1.8 dB. To test the downstream performance, we adopt splitters with 1:8 and 1:4 split ratio to simulate the scenarios of different resource requirement. A SW array is used to choose the input point of the ONU. In our experiment, we haven’t set the control channel but use a programmable control circuit through FPGA Xilinx V2. We build a model to simulate the variety of network flux. In the model, we set several cases for different network flux, so we can use different ODN structure when the network flux changed. The ODN structure is changed by adjusting the optical switches, which are driven by the transceiver port of the FPGA circuit. According to the variety, the circuit can adjust the structure of the ODN for resource efficiency. During adjustment, the performance of the system is stable.
At the ONU side, the optical signal is firstly split into two parts by a 3-dB optical coupler. One part is for downstream signal detection, and the other part is reused for 1.25-Gb/s OOK upstream. Because the performances of four-channel are the same, we only take channel of λ2 (1552.20 nm) for analysis. The optical spectra of λ2 at corresponding points in Fig. 3 are shown in Fig. 4(a)-(c). For λ2, the Tx power is 4dBm. When the network flux changes, the splitter connected to λ2 can alternate 1:4 and 1:8 split ratio. The following analyses for downstream and upstream are based on the split ratio of 1:8, which can get a worse power budget (maximum optical path loss) in the experiment. If the split ratio changes to 1:4, there would be a 3dB decrease in optical path loss.
For the downstream signal, another 3-dB optical coupler is employed to split the signal for unicast signal and DPSK signal detection respectively. One coupler output is fed into a 10-GHz photodiode (PD) for O/E conversion. The 16QAM-OFDM electrical signal is sampled by a Tektronix6804B real time digital sampling scope (TDS) for digital down-converted and offline processing. The measured bit error rate (BER) curves with and without DPSK overlaid are shown in Fig. 5 . There is about 1dB power penalty at the BER of 10−3(FEC limit) for unicast OFDM signal, which is mainly because the fiber dispersion would make the phase modulated signal partly transform into amplitude information during transmission. We can clearly see the residual amplitude information in the spectrum from Fig. 4(e). When the LO is set to 2.5 GHz, the spectra of residual amplitude information and OFDM signal will partly overlap. Comparing with the overlay-enable scenario, there is about 0.5 dB receive sensitivity deterioration for the unicast signal at the BER of 10−3. For overlay-enable downstream signals, the received sensitivity of OFDM signal at BER = 10−3 is −24.5dBm, resulting in a power budget of 28.5dB, which is same as the proposed IEEE 802.3av downstream power budget PR20. For the other part, we employ a standard DPSK balanced receiver, which consists of a one bit Mach-Zehnder delay interferometer (MZDI) demodulator and a 2.5-GHz balanced detector. The electrical spectrum after detection is illustrated in Fig. 4(f). Then the signal is passed through a 3.9-GHz LPF to suppress the frequency components higher than the DPSK bit rate, which can eliminate the crosstalk from the OFDM signal. The electrical eye diagrams and BER curves before and after transmission are shown in Fig. 6 . The power penalty is about 1 dB at the BER of 10−9 for the overlay-enable channel. Figure 6 also shows the BER curve of DPSK overlaid signal when the frequency of the LO is 2.5 GHz, where we can see that the signal cannot be recovered at all. That’s mainly because the LPF cannot filter out the OFDM signal, and the crosstalk becomes so severe as to be unable to recover the signal. Figure 6 shows both the optical eye diagram and electrical diagrams of DPSK signal. In the experiment, the bandwidth of our optical oscillograph is 60 GHz, which is much larger than the DPSK signal’s, there is more noise induced than the electrical eye diagrams. Besides, compared to the PD detection, the optical oscillograph can detected the harmonic tones of signal more completely, so the shape of the eye diagram looks squarer.
For the upstream link, the OFDM/DPSK signal is re-modulated to OOK by an IM at 1.25 Gb/s with a PRBS of length 231-1. The optical spectrum after re-modulation is shown in Fig. 4(c), where we can see the Tx power is about −13.5dBm. The spectrum of RF 16QAM-OFDM is beyond 1.25 GHz and it is out of the spectral range of the 1.25-GHz upstream OOK signal, which makes it feasible for upstream re-modulation. However, the residual amplitude information of DPSK signal at the central frequency would affect the performance of upstream signal, which is also amplitude modulated. Hence, the extinct ratio of IM is set to 10dB, which is much larger than the residual amplitude information, and the crosstalk from the residual amplitude information can be suppressed. After the transmission, a 2.5-GHz commercial PD and a 1-GHz LPF are used for direct detection. The eye diagrams and BER curves of upstream signal are illustrated in Fig. 7 . The power penalty is about 0.3 dB at the BER of 10−9 for both overlay-enable and overlay-disable scenario. For the overlay-disable channel, there is about 0.6 dB receive sensitivity deterioration compared with the overlay-enable channel. It mainly attributes to the interference from downstream signal which cannot be completely filtered out when the frequency of LO is 2.5 GHz. The receiver sensitivity of OOK upstream signal at BER=10−9 is −23.3dBm for overlay-enable channel, so the power budget is about 10dB, same as the proposed IEEE 802.3av downstream power budget PR10. Because the IM-type re-modulation cannot re-amplify the signal as RSOA does , so the power budget is less than that case. For the common single fiber system, when the downstream and upstream signals are carried on the same wavelength, the main influence comes from Rayleigh backscattering. However, our scheme employs the up-conversion OFDM signal for the downstream signal, which means the OFDM signal carried on the first-order sideband, and the upstream OOK signal carried on the central optical carrier. Thus there is little influence between the upstream signal and downstream signal because the two signals carried on different frequency in the optical domain.
We have proposed and demonstrated a novel dynamic λ-OFDMA architecture with selective DPSK multicast overlaid. The selectivity of DPSK overlaid signal is realized through an alternative local oscillator at each wavelength at the OLT. In this scheme, the downstream 4 × 10-Gb/s 16QAM-OFDM modulated unicast signal and 2.5-Gb/s DPSK multicast signal have been transmitted over 25-km SMF successfully. The receive sensitivities of unicast signals with and without DPSK overlaid are −24.9 dBm and −24.5 dBm respectively at BER of 10−3 after 25-km fiber. The power penalty for the downstream DPSK signal is 1.1 dB at BER of 10−9, while the power penalty is about 0.3 dB for the upstream signal after transmission over 25-km SMF. Our experiment work suggests the scheme a promising candidate for future optical access network.
The financial support from National Basic Research Program of China with No. 2010CB328300, National Natural Science Foundation of China with No. 61077050, 61077014, 60977046, 60932004, National High Technology 863 Research and Development Program of China with No. 2009AA01Z220, 2009AA01A347 are gratefully acknowledged. The project is also supported by the BUPT Excellent Ph.D. Students Foundation with No.CX201014, BUPT Young Foundation with No.2009CZ07, the Open Foundation of State Key Laboratory of Optical Communication Technologies and Networks (WRI) with No.2010OCTN-02.
References and links
1. G.-K. Chang, Zhensheng Jia, Jianjun Yu, Arshad Chowdhury, Ting Wang, and Georgios Ellinas, “Super-Broadband Optical Wireless Access Technologies”, in Proc. OFC, San Diego, CA, paper. OThD1 (2008).
2. C. W. Chow, C. H. Yeh, C. H. Wang, F. Y. Shih, and S. Chi, “Signal-Remodulated Wired/Wireless Access Using Reflective Semiconductor Optical Amplifier With Wireless Signal Broadcast,” IEEE Photon. Technol. Lett. 21(19), 1459–1461 (2009). [CrossRef]
3. N. U. Kim, H.-S. Lim, and M. Kang, “Fair Bandwidth Allocation Using Effective Multicast Traffic Share in TDM-PONs,” IEEE J. Lightwave Technol. 26(7), 756–767 (2008). [CrossRef]
4. J. H. Moon, K. M. Choi, and C. H. Lee, “Overlay of Broadcasting Signal in a WDM-PON”, in Proc. OFC, Anaheim, CA, paper. OTHK8 (2006).
5. G. P. Leguizamon, B. Ortega, and J. Capmany, “Advanced Subcarrier Multiplexed Label Swapping in Optical Packet Switching Nodes for Next Generation Internet Networks,” IEEE J. Lightw. Technol. 27(6), 655–669 (2009). [CrossRef]
6. M. Khanal, C. J. Chae, and R. S. Tucker, “Selective broadcasting of digital video signals over a WDM passive optical network,” IEEE Photon. Technol. Lett. 17(9), 1992–1994 (2005). [CrossRef]
7. J. Yu, M.-F. Huang, D. Qian, L. Chen, and G.-K. Chang, “Centralized Lightwave WDM-PON Employing 16-QAM Intensity Modulated OFDM Downstream and OOK Modulated Upstream Signals,” IEEE Photon. Technol. Lett. 20(18), 1545–1547 (2008). [CrossRef]
9. A. J. Lowery, L. Du, and J. Armstrong, “Orthogonal Frequency Division Multiplexing for Adaptive Dispersion Compensation in Long Haul WDM Systems”, in Proc.OFC, San Diego, CA, paper.PDP39(2006).
10. X. Liu, F. Buchali, and R. W. Tkach, “Improving the Nonlinear Tolerance of Polarization-Division-Multiplexed CO-OFDM in Long-Haul Fiber Transmission,” J. Lightwave Technol. 27(16), 3632–3640 (2009). [CrossRef]
11. D. Qian, N. Cvijetic, J. Hu, and T. Wang, “108 Gb/s OFDMA-PON with Polarization Multiplexing and Direct Detection,” J. Lightwave Technol. 28(4), 484–493 (2010). [CrossRef]
12. R. P. Giddings, X. Q. Jin, E. Hugues-Salas, E. Giacoumidis, J. L. Wei, and J. M. Tang, “Experimental demonstration of a record high 11.25Gb/s real-time optical OFDM transceiver supporting 25km SMF end-to-end transmission in simple IMDD systems,” Opt. Express 18(6), 5541–5555 (2010). [CrossRef] [PubMed]
13. D. Qian, N. Cvijetic, Y. Huang, J. Yu, and T. Wang, “100km Long Reach Upstream 36Gb/s-OFDMA-PON over a Single Wavelength with Source-Free ONUs, ” in Proc. ECOC, Austria, paper 8.5.1(2009).
14. C.-W. Chow, C.-H. Yeh, C.-H. Wang, F.-Y. Shih, C.-L. Pan, and S. Chi, “WDM extended reach passive optical networks using OFDM-QAM,” Opt. Express 16(16), 12097–12101 (2008). [CrossRef]
15. D. Qian, J. Hu, P. N. Ji, and T. Wang, “10-Gb/s OFDMA-PON for Delivery of Heterogeneous Services”, in Proc.OFC, San Diego, CA, paper.OWH4 (2009).
16. R. P. Giddings, E. Hugues-Salas, X. Q. Jin, J. L. Wei, and J. M. Tang, “Colourless Real-Time Optical OFDM End-to-End Transmission at 7.5Gb/s over 25km SSMF Using 1GHz RSOAs for WDM-PONs”, in Proc. OFC, San Diego, CA, paper.OMS4(2010).
17. L. A. Neto, A. Gharba, P. Chanclou, N. Genay, B. Charbonnier, M. Ouzzif, C. Aupetit-Berthelemot, and J. Le Masson, “High Bit Rate Burst Mode Optical OFDM for Next Generation Passive Optical Networks”, in Proc. ECOC, Italy, paper.Tu.3.B.5(2010).
18. Y. Ma, Q. Yang, Y. Tang, S. Chen, and W. Shieh, “1-Tb/s single-channel coherent optical OFDM transmission over 600-km SSMF fiber with subwavelength bandwidth access,” Opt. Express 17(11), 9421–9427 (2009). [CrossRef] [PubMed]
19. B. Liu, X. Xin, L. Zhang, J. Yu, Q. Zhang, and C. Yu, “A WDM-OFDM-PON architecture with centralized lightwave and PolSK-modulated multicast overlay,” Opt. Express 18(3), 2137–2143 (2010). [CrossRef] [PubMed]
20. A. Marc, Taubenblatt, “Optical Interconnects for High Performance Computing”, in proc.OFC, CA, USA, paper.OThH3 (2011).
21. B. S. Rho and J. W. Lim, ““Multichip Integration on a PLC Platform for 16×16 Optical Switch Module Using Passive Alignment Technique,” IEEE Trans. Adv. Packag. 30(3), 457–461 (2007). [CrossRef]
22. M. Yano, F. Yamagishi, and T. Tsuda, “Optical MEMS for Photonic Switching—Compact and Stable Optical Crossconnect Switches for Simple, Fast, and Flexible Wavelength Applications in Recent Photonic Networks,” IEEE J. Sel. Top. Quantum Electron. 11(2), 383–394 (2005). [CrossRef]
23. M. Mizukami, J. Yamaguchi, N. Nemoto, Y. Kawajiri, H. Hirata, S. Uchiyama, M. Makihara, T. Sakata, N. Shimoyama, H. Ishii, and F. Shimokawa, “128X128 3D-MEMS optical switch module with simultaneous optical paths connection for optical cross-connect systems”, in proc.Photonics in Switching, pp.1–2(2009).
24. J. L. Wei, E. Hugues-Salas, R. P. Giddings, X. Q. Jin, X. Zheng, S. Mansoor, and J. M. Tang, “Wavelength reused bidirectional transmission of adaptively modulated optical OFDM signals in WDM-PONs incorporating SOA and RSOA intensity modulators,” Opt. Express 18(10), 9791–9808 (2010). [CrossRef] [PubMed]