This paper has experimentally demonstrated and analyzed the performance of 2.5-Gb/s × 3-channel upstream transmission in electrical code divided multiplexing-orthogonal frequency division multiplexing access (ECDM-OFDM) passive optical network (PON). The colorless upstream link can be realized in ECDM-OFDM-PON. The experimental results show that the performance degradation due to optical beating interference (OBI) noise can be well suppressed in this network when the three channels adopt the same upstream wavelength. Compared with the WDM-OFDM-PON upstream signals without ECDM, the error floor shows about three orders of magnitude improvement due to the code gain when the same wavelength is used for all upstream signals.
© 2011 OSA
The appearance of new Internet services (HDTV, video telephony, etc.) has put much pressure on the existing access network. Among the proposed several technologies, passive optical networks (PON) has been considered as a promising solution due to its large throughput and quality of services (QoS) [1–4]. Recently, orthogonal frequency division multiplexing passive optical network (OFDM-PON) has been regarded as a promising solution for next generation access technology due to its robust dispersion tolerance and the flexibility on both multiple services provisioning and bandwidth allocation [5–8].
In order to realize more flexible resource allocation in OFDM-PON, it has been used to adopting multi-dimensional multiplexing in OFDM-PON. Schemes to implement the OFDM-PON include wavelength division multiplexing OFDM-PON (WDM-OFDM-PON) and time division multiplexing OFDM-PON (TDM-OFDM-PON) [9–16]. In WDM-OFDM-PON, the optical line terminal (OLT) assigns a wavelength to each optical network unit (ONU), which is a costly way for access network. Although there are some reports about realizing colorless ONU through wavelength re-modulation, it would suffer from backscattering due to optical reflections [9–11]. Besides, WDM-OFDM-PON lacks the flexibility to dynamically allocate the resources among different services and optical network units (ONUs), which reduces the efficiency of access network. TDM-OFDM-PON combines TDM and OFDM technology and enables dynamic bandwidth allocation among different services and ONUs. In TDM-OFDM-PON, the colorless ONU is still a problem due to the optical beating interference (OBI) noise which will affect the performance of upstream signals [13–15]. The previous research has used to adopt suppressed carriers transmission and coherent detection to avoid the OBI noise [14,15].
Recently, we have proposed an electrical code division multiplexing (ECDM) based OFDM-PON architecture, and it could offer bidirectional access with wavelength-based colorlessness upstream, which is intensity-modulation and direct-detection (IMDD) configuration . Nevertheless, we haven’t investigated the performance of the system in detail. In this paper, we analyze the upstream performance of ECDM-OFDM-PON with different bandwidth resource allocation, which can greatly improve the system performance when the same wavelength is used for uplink. We experimentally demonstrate a 2.5-Gb/s × 3 colorless upstream transmission over 25-km single mode fiber (SMF) successfully and further compare it with the WDM-OFDM-PON without ECDM in different channel space. The demonstration indicates a potential solution for next generation access networks.
2. System model and operating principle
Figure 1 shows the basic configuration of the ECDM-OFDM-PON. After constellation mapping, the downstream or upstream data of each ONU is assigned to the given OFDM subcarriers and code chip. The optical line terminal (OLT) controls the code chips and subcarriers allocation for each ONU according to different service demands. For the upstream signal, since one photodiode (PD) is used at the OLT, the OBI noise will occur within the receiver bandwidth and cause the performance degradation. The employment of chip code can minimize interference between different ONUs due to the code gain, while maintaining the OFDM merits of convenient multi-service and dynamical bandwidth allocation. In this architecture, the spreading spectrum with code chip is executed accompanying with the subcarrier modulation of OFDM frame in the electrical domain. The electrical ECDM-OFDM frames are converted into optical OFDM signals through intensity modulation.
The time domain ECDM-OFDM signal is given by
where m is the m th ONU, and wm is the code chip for mth ONU. The optical OFDM signal at the receiver can be expressed as
where f is the frequency of laser and γ is the modulation index. Assuming two optical signals (ONU-1 and ONU-2) are detected simultaneously, the photocurrent after the PD can be presented as
If we assume that γ1 = γ2 = γ and ignore the double frequency of photocurrent due to the low-pass photo-detection, Eq. (4) can be represented as
where f 1-f 2 denotes the central frequency of the OBI noise. After subcarriers demodulation and decoding, the received signal for one ONU can be expressed as
The first term is the data of ONU-1, the second term is the data of ONU-2 and the last term is the OBI noise, which comes from the DC components of optical sources and the spectrum itself. From Eq. (2), we can see that the second term equals zero. According to the spread spectrum theory , the maximum of the OBI noise comes to
where Gc is the code gain and BOFDM is the bandwidth of baseband OFDM signal, which is denoted 
If Gc is large enough, Nbeat can be suppressed to ensure the performance of the system.
3. Experiment and results
Figure 2 illustrates the experimental setup of upstream traffic to study the uplink performance. Three 2.5-Gb/s ECDM-OFDM signals are generated offline by Matlab programming and uploaded to the arbitrary waveform generators (AWG) 7122B with 10-bit DAC for D/A conversion. For each ONU, QPSK is used to map the PRBS sequences with word length of 215-1. 16 bit-long Walsh code chip is assigned to each data stream for spectrum spreading, and it can provide 12-dB gain theoretically. In our experiment, the FFT size is 512, and we adopt the Hamilton symmetric mapping to ensure the output is real signal as shown in Fig. 2, which can simplify the optical transmitter and reduce the cost of ONU. The length of cyclic prefix is 1/16 and training sequence is added every 60 OFDM symbols for time synchronization and channel estimation. The waveforms produced by the AWG are continuously output at a rate of 2.5-Gb/s and fed into the optical transmitter, which is consist of an external intensity modulator and a DFB laser. Each ONU transmitter launched power is set to −2 dBm. The 2.5-Gb/s × 3 upstream signals are power combined through a 4:1 coupler and then sent into the 25km single mode fiber (SMF-28) transmission link. At the OLT, a 2.5-GHz commercial PD is used as the O/E convertor. The electrical signal is then sampled by a 20-GS/s real-time scope for offline processing to recover the bit stream.
Firstly, we evaluate the performance of the WDM-OFDM-PON without ECDM in different optical channel space from 0 GHz to 50 GHz. The experimental setup of WDM-OFDM-PON without ECDM is similar to Fig. 2, and the difference is that the code chip at ONU and correlator at OLT are canceled compared with Fig. 2. Figure 3 illustrates the measured average bit rate error (BER) curves of the upstream signals from three ONUs. The corresponding error floors are about 2.26 × 10−3, 3.22 × 10−5 and 3.44 × 10−6 when the channel space is 0 GHz, 10 GHz and 50 GHz respectively. We can see that the error floor of upstream signal at 0 GHz channel space is beyond the FEC limit and cannot be recovered due to the high OBI noise. For the case of 50 GHz, the channel space seems large enough so as to neglect the effect from the OBI noise. The power penalties of the signals with different channel space can almost be ignored. Figure 3 also shows the performance comparison between ECDM-OFDM-PON and WDM-OFDM-PON with variable channel spacing. After adopting ECDM, there is about 2dB and 1.6dB receive sensitivity improved for 10GHz and 50GHz channel spaces respectively. The error floors are also improved in both 10GHz and 50GHz cases. As the number of ONUs increases, the OBI noise would also increase and the performance would even be worse. Figure 4 shows the BER performance with different ONU number and channel space. We can see that for the case of 50GHz, even the number of ONUs is 8, the performance has little deterioration, which indicates a robust resistance to OBI noise. But for the case of 10GHz, when the number of ONUs increases to 6, the OBI noise is large enough to worsen the BER performance and the error floor is observed above 10−3.
Next, we investigate the BER performances of upstream OFDM signals of ECDM-OFDM-PON. Figure 5 shows our experiment within two cases: in case-a, all the OFDM subcarriers are assigned to every ONUs, which will produce a maximum upstream rate; in case-b, only parts of the OFDM subcarriers are assigned to every ONU. The transmitter wavelengths of the three ONUs are set to 1557.37 nm. Figure 6 depicts the average BER curves of ONUs and the constellations of received signals at the OLT under the two cases. The performance deterioration due to the OBI noise is well suppressed, and there is nearly no power penalty before and after 25-km transmission. Compared with case-a, case-b gets a better receive sensitivity in the same BER performance. This is mainly because the OFDM subcarriers are not totally overlapped as in case-a, which results in a lower OBI noise from Eq. (6). The error floor of the system is almost the same as the WDM-OFDM-PON without ECDM in 50-GHz channel space. In our experiment, we just adopt three ONUs for demonstration. Similarly, the OBI noise would enlarge as the number of ONUs increases. However, the OBI noise can still be suppressed through extending the length of code chip. The complexity scaling is proportioned to N × (N-1), where N is the chip length, and it mainly attributes to the shift register executing the encoding part. For upstream signals with fixed bandwidth, both the chip length and number of ONUs will affect the OBI noise and system performance in ECDM-OFDM-PON. According to the Q-factor definition  and Eqs. (3)-(7), the Q-factor of the upstream can be expressed as
where Nonu is the number of ONUs and N is the chip length. Equation (9) indicates a number of tolerable OBI noise sources and the corresponding chip length.
We have experimentally demonstrated a 2.5-Gb/s × 3 colorless upstream transmission in ECDM-OFDM-PON. We investigated the BER performances of WDM-OFDM-PON without ECDM system in different channel space as well as ECDM-OFDM-PON. The experimental results show that the ECDM-OFDM-PON can improve the error floor of the system by three orders of magnitude when the same wavelength is used for all upstream signals. This paper also suggests some guidance on number of ONUs and chip length for ECDM-OFDM-PON.
The financial support from National Basic Research Program of China with No. 2010CB328300, National Natural Science Foundation of China with No. 61077014, 61077050, 60932004, BUPT Young Foundation with No. 2009CZ07 are gratefully acknowledged. The project is also supported by the Fundamental Research Funds for the Central Universities and the open foundation of state key laboratory of optical communication technologies and networks (WRI) with No. 2010OCTN-02.
References and links
1. P. P. Iannone and K. C. Reichmann, “Optical access beyond 10 Gb/s PON,” in 2010 36th European Conference and Exhibition on Optical Communication (ECOC) (2010), paper.Tu.3.B.1, pp. 1–5.
2. M. Dueser, “Optical network architectures,” in Optical Fiber Communication Conference, OSA Technical Digest (CD) (Optical Society of America, 2011), paper OMN1.
3. M. Cvijetic, “Advanced Technologies for Next-Generation Fiber Networks,” in Optical Fiber Communication Conference, OSA Technical Digest (CD) (Optical Society of America, 2010), paper OWY1.
4. G. Chang, Z. Jia, J. Yu, A. Chowdhury, T. Wang, and G. Ellinas, “Super Broadband Optical Wireless Access Technologies,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference, OSA Technical Digest (CD) (Optical Society of America, 2008), paper OThD1.
5. J. L. Wei, C. Sánchez, R. P. Giddings, E. Hugues-Salas, and J. M. Tang, “Significant improvements in optical power budgets of real-time optical OFDM PON systems,” Opt. Express 18(20), 20732–20745 (2010). [CrossRef] [PubMed]
6. Y.-M. Lin and P.-L. Tien, “Next-generation OFDMA-based passive optical network architecture supporting radio-over-fiber,” IEEE J. Sel. Areas Comm. 28(6), 791–799 (2010). [CrossRef]
7. J. G. Lin Chen, J. G. Yu, J. Shuangchun Wen, Z. Lu, M. Dong, Huang, and G. K. Chang, “A novel scheme for seamless integration of ROF with centralized lightwave OFDM-WDM-PON system,” J. Lightwave Technol. 27(14), 2786–2791 (2009). [CrossRef]
8. W. Wei, C. Wang, J. Yu, N. Cvijetic, and T. Wang, “Optical orthogonal frequency division multiple access networking for the future internet,” J. Opt. Commun. Netw. 1(2), A236–A246 (2009). [CrossRef]
9. J. M. Tang, P. M. Lane, and K. A. Shore, “High-speed transmission of adaptively modulated optical OFDM signals over multimode fibers using directly modulated DFBs,” J. Lightwave Technol. 24(1), 429–441 (2006). [CrossRef]
10. 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]
12. C. W. Chow, G. Talli, A. D. Ellis, and P. D. Townsend, “Rayleigh noise mitigation in DWDM LR-PONs using carrier suppressed subcarrier-amplitude modulated phase shift keying,” Opt. Express 16(3), 1860–1866 (2008). [CrossRef] [PubMed]
13. 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.
14. N. Cvijetic, D. Qian, J. Hu, and T. Wang, “44-Gb/s/λ Upstream OFDMA-PON Transmission with Polarization-Insensitive Source-Free ONUs,” in Optical Fiber Communication Conference, OSA Technical Digest (CD) (Optical Society of America, 2010), paper OTuO2.
15. N. Cvijetic, D. Qian, and J. Hu, “100 Gb/s optical access based on optical orthogonal frequency division multiplexing,” IEEE Commun. Mag. 48(7), 70–77 (2010). [CrossRef]
17. A. W. Lam and S. Tantaratana, Theory and Application of Spread-Spectrum Systems (IEEE, Piscataway, NJ, 1994).
19. G. P. Agrawal, Fiber-Optic Communication Systems (Wiley-Interscience, New York, 1997).