Open Access
Add to BibSonomyAbstract
We introduce a novel architecture for next generation passive optical network (PON) base on the Single-carrier Frequency Division Multiple Address (SC-FDMA) technique. Both downstream and upstream SCFDMA-PON transmissions (5 Gb/s total, 2.5 Gb/s for each user) are experimentally demonstrated over 22.2 km standard single mode fiber and an additional simulated 1:32 optical splitter. We also test the tolerance range of the synchronization error and prove it matches the cyclic prefix period in our scheme, which means the packet transmission accuracy from different optical network units can be relaxed in the upstream.
©2010 Optical Society of America
1. Introduction
With the rapid development of various communication services in the last decades, the customers are no longer only satisfied by voice telephony, broadcast television and radio. They are increasingly interested in new kinds of multi-media communications, such as fast Internet communication, high definition multimedia television (HDTV) and fast peer-to-peer file transfer. The passive optical network (PON) is promoted as a promising solution for future broadband access networks. It can provide different customers with simultaneous delivery of multi-services over a common network platform [1–3]. The standardization process aiming at defining a symmetric 10 Gb/s downstream/upstream PON specification has been worked on. And due to the growing bandwidth demand in the future, the next-generation PON technologies at the traffic higher than 10 Gb/s is also widely discussed now.
To support multiple users in one common network, several access technologies have been proposed, such as Time Division Multiple Access (TDMA), Wavelength Division Multiplexing (WDM), Optical Code Division Multiplexed Access (O-CDMA), and Orthogonal Frequency Division Multiplexing Access (OFDMA) [1,4]. Among them, the OFDMA scheme is impressing for its highest 108-Gb/s single wavelength downstream/upstream data rate in PON systems [5]. In addition, the digital signal processing (DSP)-based OFDMA-PON can provide other highly desirable features such as high spectral efficiency, high tolerance against various fiber dispersion effects and extreme flexibility on both multiple services access and dynamic bandwidth allocation [6–8]. However, due to its high peak-to-average ratio (PAPR), the OFDMA technique requires highly linear power amplifiers to avoid excessive intermodulation distortion. This would increase the cost of the optical architecture, especially for the optical network unit (ONU) side.
In this paper, we propose a novel PON architecture employing the Single-carrier Frequency Division Multiple Address (SC-FDMA) technique. The SC-FDMA is a modified form of OFDMA and is currently employed for the uplink multiple access scheme in the Long Term Evolution (LTE) of cellular systems by the Third Generation Partnership Project (3GPP) [9,10]. Due to its inherent single carrier transmission characteristics, it has lower PAPR under similar throughput performance and overall complexity compared with OFDMA [11,12]. In this paper, we show by experiments that SC-FDMA is feasible for both downstream and upstream PON transmission.
2. Technique principle
As a modified form, the baseband DSP method of the SC-FDMA has much in common with that of the OFDMA. Figure 1 shows the transmitter and receiver DSP block diagram for SC-FDMA. We can see that the major differences between OFDMA and SC-FDMA are the presences of the discrete Fourier transform (DFT) in the SC-FDMA transmitter and the inverse DFT (IDFT) in the SC-FDMA receiver. For this reason, SC-FDMA is sometimes referred to as DFT-spread OFDMA. At the transmitter, the first step to generate the SC-FDMA symbol is to perform an M-point DFT to produce the frequency domain representation of mapped Quadrature Amplitude Modulation (QAM) or Phase Shift Keying (PSK) single-carrier signals. It then encapsulates each of the M-point DFT outputs to one of the N (N> M) orthogonal subcarriers that can be transmitted. After subcarrier mapping, an IDFT transforms the subcarriers to a complex time domain signal. Before the signal is transmitted, cyclic prefix (CP) is inserted in order to provide a guard time to prevent inter-block interference. At the receiver, after CP is removed from each block, the N-point DFT transforms the signals into frequency domain and channel equalization is performed. Unlike OFDMA scheme in which decision is performed in the frequency domain, the equalized SC-FDMA signal is transformed into time domain by the M-point IDFT for further decision.
We can see that subcarrier mapping enables spectrum division for different terminals and is essential for SC-FDMA. The mapping can be localized or distributed [12]. Figure 2 shows a localized SC-FDMA scheme for a typical PON architecture. In OFDMA each ONU uses a set of subcarriers to transmit or receive its data. In SC-FDMA, the ONUs employ single-carrier transmission, but each single carrier is frequency-domain shifted to occupy a specific pre-assigned part of the whole available bandwidth. Since they transmit the subcarriers sequentially rather than in parallel, the envelope fluctuations in the transmitted waveform of SC-FDMA signals is reduced and the burden of linear amplification at the user is mitigated. For upstream transmission, each ONU should set unwanted subcarriers to zeroes during the subcarrier mapping. Due to the DFT guaranteed orthogonality, if all the reference clocks of different ONUs are synchronized, the optical line terminal (OLT) can receive data from all ONUs simultaneously without inter-subcarrier interference after standard single mode fiber (SSMF) transmission.
Moreover, as shown in Fig. 3 , if the traffic is organized with SC-FDMA frames, each of which consists of multiple SC-FDMA symbols, the resource elements can be two-dimensional in both frequency and time domains. Various multiple services for different users can be encapsulated into the given subcarriers and time slots according to the proper frequency/time domain schedule. So the resource allocation in SCFDMA-PON is as flexible as it is in OFDMA-PON [13].
To protect the legacy network investments, there is interest in providing next-generation PON primarily over TDM techniques, such as 10G-PON and 10GE-PON. But the TDM-based PON techniques ask for complex scheduling algorithms and framing technology, and are highly sensitive to packet latency in the upstream [6]. In the OFDMA-based PON or the SCFDMA-PON we proposed in this paper, since the information is transmitted in parallel in the frequency domain and CPs are inserted between the transmission data blocks to provide a period of guard time, the requirement of packet upstream transmission accuracy is much relaxed. As shown in Fig. 4 , take the two-ONUs-upstream case as an example, we fix the ONU-1 upstream as the constant and let ONU-2 upstream transmission ahead or delay. Since the CP is a copy of part of the data block and according to the circular convolution characteristic of DFT, the demodulation output remains correct if the synchronization error stays within from –CP/2 to CP/2. So the tolerance range of the packet latency in our SCFDMA-PON is the period of CP.
In cellular applications, SC-FDMA is not recommended as the downstream technique, because OFDMA performs better in the presence of severe multipath signal propagation. The immunity to multipath derives from the fact that an OFDMA system transmits information on multiple orthogonal frequency carriers, which is more robust against frequency selective fading than the single-carrier approach. But optical fibre channel differs from wireless channel since there is no frequency selective fading caused by multipath effects. In this paper, we experimentally verify the feasibility of SCFDMA-PON with both downstream and upstream SC-FDMA transmission.
3. Experiment setup
Figure 5 shows the experimental setup to validate the SCFDMA-PON architecture. The baseband SC-FDMA signal is generated with three times up-sampling and up-converted to 2.5 GHz by digital I-Q modulation in Matlab. The FFT size is 256 and from which 204 subcarriers are used for data transmission. The CP size is 16 and QPSK is used for constellation mapping. The roll-factor of the pulse shaping filter we used is 0.08 which is suitable in our experiment to prevent out-of-band interference. The generated waveform is uploaded into a Tektronix AWG7122B whose waveforms are continuously output at a sample rate of 10 GS/s (8 bits DAC), the total bit rate for two ONUs is 5 Gb/s (2.5 Gb/s for each ONU). An intensity Mach-Zehnder modulator (MZM) is utilized to convert the SC-FDMA signal to double-side-band (DSB) optical signal. The optical distribution network is emulated with 22.2 km SSMF, a 10-dB fixed optical attenuator, a variable optical attenuator (VOA) and a 1:2 splitter. The optical DSB signal is converted to electrical RF signal by a photodiode and then is amplified before sampled by a real-time digital storage oscilloscope (Tektronix DPO72004B) at the sampling rate of 25 GS/s. The sampled data are decoded in offline process.
Fig. 5 Experimental setup for downstream and upstream SCFDMA-PON. (LD: laser doxide, EA: electrical amplifier, VOA: variable optical attenuator, PD: photodiode).
For downstream traffic, data from OLT can be a whole single-carrier band or multiple single-carrier bands assigned to different ONUs. In the experiment, the latter case is preferred and two single-carrier bands for two ONUs are generated independently and combined in Matlab. For upstream traffic, the two output channels of AWG independently modulate two intensity MZMs to emulate two ONUs’ upstream signals. They are combined in the 1:2 splitter for uplink SSMF transmission. The wavelength of downlink laser is 1550 nm. But in the uplink, since the optical carriers from different ONUs are mixed in the same photodiode (PD), using the same upstream wavelength will induce severe received power fluctuation due to optical interference effect. It will degrade the uplink performance greatly. So different wavelengths (1550 nm for ONU-1 and 1557 nm for ONU-2) are chosen in the uplink, which is the same case in [14]. To achieve colourless ONU in the uplink, the similar coherent-detection technologies used in OFDMA-PON [15] can also be introduced in SCFDMA-PON.
4. Results and discussion
Figure 6 shows the received downstream and upstream SC-FDMA signal spectrum. Since the reference clock at the two ONUs may not be synchronized in the experiment, as studied in [16], a guard band should be put between the two ONUs’ upstream signal to minimize the possible interband interference due to the lack of frequency orthogonality. To overcome this problem, frequency offset pre-compensation can be used here to keep the reference clocks at each ONUs synchronized with each other as it is shown in OFDMA multiuser uplink systems [17]. In our experiment, a guard band of 200 MHz is chosen for both downstream and upstream to separate and distinguish the two ONUs’ signals. It should be pointed out that the optimization of guard band is not considered here, which is necessary when more ONUs are served. The 2.6 ~4 GHz frequency band is assigned to ONU-1, while the 1 ~2.4 GHz frequency band is assigned to ONU-2.
Fig. 6 Signal spectrums. (a) downstream for ONU-1&2; (b) simultaneous ONU-1&2 upstream of SC-FDMA signals; (c) single ONU-1 upstream of SC-FDMA signals; (d) single ONU-2 upstream of SC-FDMA signals.
The received bit error rate (BER) performance for downstream and upstream traffic shown in Fig. 7 and Fig. 8 are calculated from the measured error vector magnitude (EVM) [18]. In the downstream traffic, the transmitter power is 7.5 dBm. For ONU-1 and ONU-2, a BER of 10−3 can be achieved with the received optical power at about −18.3 dBm after downlink transmission, where the loss budget is 25.8 dB. This represents an excellent transmission performance where a BER of 1 × 10−9 or less can obtained with the help of forward error correction (FEC) technology. The CP time we used in the experiment is 4.8 ns, which is long enough to compensate the chromatic dispersion for fiber distance exceeding 1000 km [19]. So the power penalty introduced by downlink SSMF transmission can be neglected compared to the back-to-back (BTB) case. In the upstream traffic, the transmitter power of ONU-1 and ONU-2 is 7.5 dBm and 12 dBm. Under the combined transmission, the received optical power for ONU-1 and ONU-2 at a BER of 10−3 are about −19.3 dBm and −16.2 dBm, respectively. The loss budgets are 26.8 dB for ONU-1 and 28.2 dB for ONU-2. The power penalty compared BTB case is too small to be detected. The different received sensitivities between the two ONUs come from the different MZM modulation intensity and linearity. Since the inherent loss of 20-km SSMF and 1:32 splitter is about 19 dB, the additional power budget for ONU-1 and ONU-2 are 7.8 dB and 9.2 dB. So the maximal allowable difference in the uplink transmission distance between the two ONUs is about 46 km, if the dynamic range of the OLT receiver is enough and the accuracy of synchronization is guaranteed. To evaluate the interference between the two ONUs, the BER performances of single upstream transmission for each ONU are also tested. The interference for ONU-1 can be neglected while there is about 0.3-dB power penalty at a BER of 10−3 for ONU-2 between the single and combined transmission. This penalty is attributed to noise floor of ONU-1 upstream in the frequency band of ONU-2.
The tolerance range of the synchronization error in the experiment is also measured. If the synchronization position is chosen according to the ONU-1’s upstream, the BER performance of both the two ONUs is tested and shown in Fig. 9 when the ONU-2’s upstream transmission position is changed manually. The received optical powers for ONU-1 and ONU-2 in this test are −16.2 dBm and −13.6 dBm respectively to keep both of the two ONUs in good BER performance (10−5~10−6). As illustrated in Fig. 9, the BER performance of ONU-2 remains almost the same within the synchronization error range from −24 to 24 while it is degraded sharply outside the range. Since the CP size is 16 and the up-sampling rate is 3 in our system, the tolerance range of the synchronization error (48 samples under 10 GS/s sampling) in our experiment matches well with the deduction we introduced in the technique principle.
Fig. 9 The BER performance under different synchronization errors. (Take the ONU-1 as the synchronization criterion).
5. Conclusions
We have proposed a novel PON architecture employing the Single-carrier Frequency Division Multiple Address (SC-FDMA) technique. The first symmetric 5-Gb/s optical SCFDMA-PON traffic (2.5 Gb/s for each ONU) is experimentally demonstrated over 22.2 km SSMF followed by an additional simulated 1:32 optical splitter. The tolerance range of the synchronization error is also measured and it matches the CP period very well. The experiment has proven the feasibility of the proposed SCFDMA-PON architecture as a flexible, high-speed, and cost-efficient access network for future broadband access and delivery of heterogeneous services.
Acknowledgment
This work was supported by National Basic Research Program of China (973 Program, No. 2010CB328201 and 2010CB328202), National Natural Science Foundation of China (NSFC, No. 60907030, No. 60877045, No. 60932004 and No. 60736003), and National Hi-tech Research and Development Program of China (No. 2009AA01A345).
References and links
1. T. Koonen, “Fiber to the Home/Fiber to the Premises: What, Where, and When?” Proc. IEEE 94(5), 911–934 (2006). [CrossRef]
2. L. G. Kazovsky, W.-T. Shaw, D. Gutierrez, N. Cheng, and S.-W. Wong, “Next-Generation Optical Access Networks,” J. Lightwave Technol. 25(11), 3428–3442 (2007). [CrossRef]
3. R. Davey, J. Kani, F. Bourgart, and K. McCammon, “Options for Future Optical Access Networks,” IEEE Commun. Mag. 44(10), 50–56 (2006). [CrossRef]
4. D. Qian, J. Hu, J. Yu, P. N. Ji, L. Xu, T. Wang, M. Cvijetic, and T. Kusano, “Experimental Demonstration of a Novel OFDM-A Based 10 Gb/s PON Architecture,” in Proc. 33th European Conf. on Opt. Commun. (ECOC 2007), paper Mo 5.4.1, 2007.
5. D. Qian, N. Cvijetic, J. Hu, and T. Wang, “108Gb/s OFDMA-PON with Polarization Multiplexing and Direct-Detection,” in Optical Fiber Communication Conference, OSA Technical Digest (CD) (Optical Society of America, 2009), paper PDPD5.
6. D. Qian, J. Hu, P. Ji, T. Wang, and M. Cvijetic, “10-Gb/s OFDMA-PON for Delivery of Heterogeneous Services,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference, OSA Technical Digest (CD) (Optical Society of America, 2008), paper OWH4.
7. P. Tien, Y. Lin, and M. C. Yuang, “A Novel OFDMA-PON Architecture toward Seamless Broadband and Wireless Integration,” in Optical Fiber Communication Conference, OSA Technical Digest (CD) (Optical Society of America, 2009), paper OMV2.
8. 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), 12096–12101 (2008), http://www.opticsinfobase.org/abstract.cfm?URI=oe-16-16-12096. [CrossRef] [PubMed]
9. 3rd Generation Partnership Project, “Physical Layer Aspects for Evolved Universal Terrestrial Radio Access (UTRA),” (2006), http://www.3gpp.org/ftp/Specs/html-info/25814.htm.
10. H. Ekström, A. Furuskär, J. Karlsson, M. Meyer, S. Parkvall, J. Torsner, and M. Wahlqvist, “Technical Solutions for the 3G Long-Term Evolution,” IEEE Commun. Mag. 44(3), 38–45 (2006). [CrossRef]
11. D. Falconer, S. L. Ariyavisitakul, A. Benyamin-Seeyar, and B. Eidson, “Frequency Domain Equalization for Single-Carrier Broadband Wireless Systems,” IEEE Commun. Mag. 40(4), 58–66 (2002). [CrossRef]
12. H. G. Myung, J. Lim, and D. J. Goodman, “Single Carrier FDMA for Uplink Wireless Transmission,” IEEE Trans. Veh. Mag. 1(3), 30–38 (2006). [CrossRef]
13. 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]
14. D. Qian, J. Hu, P. N. Ji, and T. Wang, “10.8-Gb/s OFDMA-PON Transmission Performance Study,” in National Fiber Optic Engineers Conference, OSA Technical Digest (CD) (Optical Society of America, 2009), paper NME6.
15. 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. 35th European Conf. on Opt. Commun. (ECOC 2009), paper 8.5.1, 2009.
16. Q. Yang, W. Shieh, and Y. Ma, “Guard-band influence on orthogonal-band-multiplexed coherent optical OFDM,” Opt. Lett. 33(19), 2239–2241 (2008). [CrossRef] [PubMed]
17. J.-J. van de Beek, P. O. Borjesson, M.-L. Boucheret, D. Landstrom, J. M. Arenas, P. Odling, C. Ostberg, M. Wahlqvist, and S. K. Wilson, “A Time and Frequency Synchronization Scheme for Multiuser OFDM,” IEEE J. Sel. Areas Comm. 17(11), 1900–1914 (1999). [CrossRef]
18. V. J. Urick, J. X. Qiu, and F. Bucholtz, “Wide-band QAM-over-fiber Using Phase Modulation and Interferometric Demodulation,” IEEE Photon. Technol. Lett. 16(10), 2374–2376 (2004). [CrossRef]
19. W. Shieh, W. Chen, and R. S. Tucker, “Polarisation mode dispersion mitigation in coherent optical orthogonal frequency division multiplexed systems,” Electron. Lett. 42(17), 996–997 (2006). [CrossRef]
