Open Access
Add to BibSonomyAbstract
This paper proposes a novel wavelength division multiplexing-orthogonal frequency multiple access (WDM-OFDMA) union passive optical network (uni-PON) architecture with dynamic resource allocation and variable rate access. It can offer an infrastructure with different access solutions. According to the quality of service (QoS) requirement of different services, the optical local terminal (OLT) can dynamically assign different resources as well as the access rates to different services. An experiment has been demonstrated with 4 wavelengths achieving combined signal at 109.92-Gb/s. A physical-layer adaptive algorithm is employed for the resource allocation and variable rate access. The different services with different resource allocations and variable access rates are also demonstrated in the experiment.
©2012 Optical Society of America
1. Introduction
Rapid growth of the Internet traffic and services such as HDTV, multimedia conference and the future passive optical network (PON) requires high data rate and broadband connectivity as well as flexible scalability [1–4]. Besides, due to the different access environments, the network also needs to highlight its key role in delivering high bandwidth services to any type of users. The full service access network (FSAN) NG-PON2 White Paper Drafts has proposed several key issues for future PONs: 40-100 Gb/s capacity, 256-1000 sustained ONUs, access distance larger than 60km and compatibility with wireless signal. However, due to the fixed initial architecture of the current PONs such as Ethernet PON, the services provider cannot meet the future demands of capacity and bandwidth etc [5–7]. Several optical access technologies have been proposed to meet the demands of future access networks, including wavelength division multiplexing (WDM) PONs, WDM/TDM PONs and orthogonal frequency division multiplexing access (OFDMA) PONs [8–13]. However when the access speed upgrades to 40-Gb/s and beyond, the high-speed burst receiver and timing becomes an intractable problem for TDM-PON. Although WDM-PON is with abundance of capacity, it requires multiple transceivers which will greatly increase the system cost. Moreover each wavelength is fixed for one ONU in WDM-PON which causes a rough granularity and wastes the resource. Nevertheless above architectures are focused on the home/enterprise access solutions which ignore the connectivity between the radio remote units (RRUs) as well as the indoor high speed mobile solutions. To achieve all-service access network and reduce the system cost, the service providers are strongly expected to investigate a union PON for next generation PONs.
Recently, China Mobile and its industrial partners such as IBM, ZTE and Intel have organized a cooperation project called Cloud-Radio Access Network (C-RAN), in which a concept of union PON (uni-PON) has been proposed for the future access networks [14]. It involves both baseband and radio access signals which are transmitted through fiber to RRU for subsequent transmission through antenna. In the uni-PON, all the data processing is concentrated at the OLT with the help of the cloud computing which improves the system efficiency as well as reduces the cost. The uni-PON cannot only provide all-services access for any type of users, but also cover high-frequency indoor networking and the connectivity of RRUs. Moreover it can support flexible access featuring a dynamic bandwidth allocation and a multi-rate adjustment according to different quality of service (QoS) or access environments. OFDM modulation is an attractive technology to realize the above characteristics in the physical layer. On the other hand, due to the emergence of huge broadband services, PON beyond Gigabits has received high attention across the world. With record of 1.92-Tb/s rate and subcarriers sharing, OFDM technology has also emerged as an attractive candidate for next generation high speed access [13, 15]. OFDM characteristics such as high spectral efficiency, gigabit broadband and resistance to chromatic dispersion and polarization mode dispersion enables a network to accommodate higher bit rates [15–18].
In this paper, we propose and demonstrate a novel WDM-OFDMA uni-PON architecture for the future broadband access, which can offer both ONU and RRU access with dynamic resource allocation and variable rate. In the proposed scheme, we further adopt an adaptive algorithm by leading symbol at physical-layer for the dynamic allocation and variable access rate. In our experiment, the access rate can vary from 13.74 Gb/s to 27.48 Gb/s with a total 5GHz bandwidth per wavelength. A total bit rate of 109.92-Gb/s consisting of three services with 4 wavelengths has successfully been reached in the experiment.
2. System configuration
Figure 1 illustrates the proposed configuration of the uni-PON. The topology consists of optical local terminal (OLT) and several ONUs/RRUs which are located at different premises such as residential or corporate. The data traffic is carried on different wavelengths and OFDM subcarriers. The OLT can dynamically assign different wavelengths and OFDM subcarriers to different ONUs/RRUs for different services which include voice/video signal, point to point (P2P) and Wi-Fi signals. At the OLT, each CW light is sliced into three optical carriers through a Mach-Zenhder modulator (MZM) and a radio frequency (RF) clock. One is the central carrier and the other two are first-order sidebands which are named as couple carriers. Then an optical filter (OF) is adopted to separate the central optical carrier and couple carriers for OFDM data modulation as shown in Fig. 1. The central carrier is reserved for the baseband services and the couple carriers are reserved for radio services which utilize the radio over fiber (ROF) technology [19]. According to the types of services or receivers, the information would be added to either or both kinds of optical carriers. Due to the released 7 GHz license-free band, 60 GHz wireless communication has attracted much attention and the next generation Wi-Fi access is also working at 60 GHz [20]. Thus in our following experiment, we choose 60GHz as the millimetre wave (mm-wave) carrier of radio service which means the RF clock is 30GHz. Then the N wavelengths are combined through a multiplexer (Mux) for transmission. In Ref [21], it also adopts MZM and IL for the OFDM baseband and ROF generation. However, the generated four optical carriers are used for wired and radio data where the radio signal is actually added onto the central baseband optical carrier. Furthermore, the network architecture in Ref [21]. is a static one.
Fig. 1 The proposed uni-PON architecture (CW: continues wave laser; MZM: Mach-Zenhder modulator; IM: intensity modulator; MUX: multiplexer; OF: optical filter; FWB: flexible wavelength blocker; AWG: arrayed waveguide grating; SW: switch; RN: remote node).
When the traffic is transmitted to the remote node (RN), a flexible wavelength blocker (FWB) is adopted to route the optical carriers for different ONUs/RRUs according to their requirements as illustrated in Fig. 1. There we group the ONUs/RRUs into three types: type A contains only radio services (wireless signal), type B contains only baseband services and type C includes both services. The FWB is consisting of an arrayed waveguide grating (AWG), an optical switcher (SW) and some optical couplers. In Fig. 1, the input signal of FWB is a group of optical carriers spaced around 33GHz, and the total number of the optical carriers is 3 × N, where N equals the number of wavelength channel. The input optical carriers are firstly demultiplexed by the AWG and then fed into the SW with 3 × N inputs and 3 × N outputs. According to the RRU/ONU requirement, the SW can deliver different optical carriers to them. The SW can be controlled by the programmable module and the control information comes from the OLT through a control channel. The granularity of the FWB depends on the channel space of AWG, which is 33GHz in our scheme. The optical couplers after the SW are used to combine the chosen optical carriers. For example, in type A case, the AWG separates all the optical carriers and the SW will choose the dedicated couple carriers for RRU to generate the wireless signal. The ONUs or RRUs extract their own information from different OFDM subcarriers.
In order to realize the dynamic bandwidth allocation and variable access rate, we employ a physical-layer adaptive algorithm for the OFDM signal which is shown in Fig. 2 . To illustrate the algorithm, we assume there are three kinds of services and four kinds of m-QAM mapping formats. Figures 2(a) and 2(b) show the transmitter and receiver diagrams for the dynamic allocation which is realized through the leading symbol. The structure of the leading symbol is shown in Fig. 2(c) and it is also with OFDM modulation format. The subcarriers of leading symbol are consisting of two parts: resource label and speed label. The resource label is used to record the bandwidth allocation for different services; the speed label is used to record the different access bit rates for different services, which are realized through different mQAM mappings. The speed label contains several subcarriers which represent different QAM mapping formats as shown in Fig. 2(c). When one QAM mapping format is chosen (e.g. 16QAM), the power value of the corresponding subcarrier will be set to the root mean square of 16QAM and the other subcarriers would be set to zero. At the transmitter, the resource label stores the starting subcarrier index of each service and the speed label stores the mQAM mapping adopted for each services. At the receiver, the ONU/RRU extracts the subcarrier and mQAM mapping information from the leading symbol. When the bandwidth and access rate are changed, the OLT and ONU can automatically adapt the variety through the leading symbol which is transparent to the media access control (MAC) layer.
Fig. 2 The block of adaptive algorithm: (a) transmitter; (b) receiver; (c) structure of leading symbol (P/S: parallel to serial; IFFT: inverse fast Fourier transform).
3. Experimental setup and results
The experimental setup is shown in Fig. 3 , where we adopt four wavelength channels for the demonstration. At the OLT, four DFB lasers at 1549.32nm, 1550.12 nm, 1550.92nm and 1551.72nm (λ1 to λ4) are employed as the light sources which satisfy the ITU-T standard grid. In Ref [22], a four-band OFDM-PON with variable access rate is proposed which realized 40Gb/s wired access with 10 GHz bandwidth. However, the four bands are electrical signal bands and it lacks physical-layer adaptive algorithm for variable rate. The four bands are multiplexed through electrical I/Q modulation and the rate for each electrical band is fixed at the OLT. In our experiment, we adopt four optical wavelengths for both wired and wireless access and a physical-layer adaptive algorithm is proposed for automatic allocation of the bandwidth and access rate. To reduce the system complexity and cost, we adopt one electrical band with total bandwidth of 5 GHz for each wavelength. It can be extended to more bands if higher rate is required. At the OLT, the four wavelengths are firstly sliced by a 30 GHz radio frequency (RF) clock to generate the optical carriers and the optical spectrums are shown in Fig. 4(a) where we can see that the center frequency of mm-wave optical carrier for the radio signal is 60 GHz. OFDM is a kind of an analog signal and the ROF is also analog. A 100 GHz inter-leaver (IL) is adopted to separate the odd and even wavelength channels. After that, two 25 GHz ILs are employed to divide the central carriers and couple carriers. The central carriers are used to carry the baseband data and the couple carriers are used to carry the radio data. They are both directly modulated through the IMs which are working in the linear area at 1.7V with half-wave voltage of 3.5V. The corresponding optical spectra after modulation are shown in Figs. 4(b)-4(c).
Fig. 3 Experimental setup for the proposed uni-PON architecture (IL: interleaver; SMF: single mode fiber; LPF: low pass filter; TDS: real time domain sampling scope)
Fig. 4 The corresponding optical spectra in Fig. 3 with resolution bandwidth of 0.02nm: (a) after the MZM; (b) after IM for radio signal; (c) after IM for baseband signal; (d) after EDFA; (e) before Rx-1; (f) before Rx-2.
The OFDM signals are generated offline by MATLAB program and then uploaded to the commercial 10Gs/s arbitrary waveform generator (7122B) with 8 bit digital-analog converter (DAC). The OFDM signal is digitally up-converted to produce a real signal. Two arbitrary waveform generators are used to generate the transmitted signals for even and odd wavelength channels respectively. The I and Q ports are used to produce the baseband and radio signals respectively. The RF spectra of the signals are shown as insets in Fig. 3. The 3.5GHz baseband signal is modulated onto the central carrier and directly detected for wired users. The 1.5GHz radio signal is modulated onto the 60-GHz spaced couple carriers and optically up-converted to 60GHz wireless signal at the RRU. A total number of 512 subcarriers are utilized which includes 27 blank subcarriers and 16 pilot subcarriers. The cyclic prefix of OFDM data symbol is 1/16 symbol time. In the experiment, we divide the whole OFDM subcarriers into three services and each service can use different QAM mapping formats from 8QAM to 64QAM as shown in Fig. 2. The net data rate on each wavelength can be changed from 13.74 Gb/s (9.62 Gb/s for baseband and 4.12 Gb/s for radio) to 27.48 Gb/s (19.24 Gb/s for baseband and 8.24 Gb/s for radio) and from 54.96 Gb/s to 109.92 Gb/s for four wavelengths. The leading symbol is attached to every 50 OFDM data symbols, which indicates the subcarrier allocation and constellation mapping formats of the three assumed services. Originally, each service is designed to have 1.67 GHz bandwidth and 64QAM-OFDM modulation to achieve 109.92 Gb/s transmission rate. If there is any change in resource allocation or access rate requirements, the OLT can choose an appropriate subcarriers and constellation mapping through the leading symbol which guarantees a dynamic and variable characteristics of the uni-PON. After modulation, the four wavelength channels are combined by a multiplexer (Mux) and then sent into the 25km fiber link and the average optical power is set to 5dBm for each wavelength channel. Figure 4(d) shows the optical spectrum of combined signal.
In the experiment, the ONUs/RRUs are of three types: type A, type B and type C. After transmission, the FWB is used to separate the optical carriers from the OLT and delivered the corresponding optical carriers to different RRU/ONUs through another 5km SMF: the central carrier for wired access users and the couple carriers for wireless access users. The FWB is consisting of an AWG and a 12x12 SW. The SW is controlled through programmable module in the experiment. The granularity of the FWB is 33GHz. The total insert loss of the element is about 4.5dB (1.3dB for SW and 3.2dB for AWG). If the number of wavelength increases, AWG and SW with larger scale would be needed and it will increase the system complexity and cost.
For the type A case, a 60 GHz PD is adopted for the generation of the wireless OFDM signal, which utilizes the optical up-conversion theory [23]. The generated RF signal centered at 60 GHz can be set into the air through antenna directly. At the users’ terminal, the RF signal can be down-converted with a 60 GHz local oscillator (LO). The down-converted OFDM signal is fed into a real time domain sampling scope (TDS) with 20Gs/s sample rate to capture the waveform for offline DSP processing. The demodulation process includes the software down-conversion, synchronization, leading symbol extraction, FFT processing, equalization and QAM de-mapping. For the type B case, the optical OFDM signal is directly detected by a photodiode (PD) with 3 dB bandwidth of 5 GHz. Then the detected electrical wired signal is also sampled by a TDS for offline DSP processing. For the type C case, additional 25 GHz IL is employed to separate the central carrier and couple carriers. The optical spectra of λ1 after separation are shown in Figs. 4(e) and 4(f). After separation, the detections of couple carriers and central carrier are same as of type A and type B respectively. Considering that both wired and wireless signals are included in type C case, we have measured the performance of this case for demonstration. We have measured OFDM signals with different QAM mapping varying from 8QAM to 64QAM, the total measured bits are 0.7 to 1.41 million each time.
Figure 5 illustrates the measured BER curves for both baseband and radio 64QAM-OFDM signals at back to back (b2b) and 30 km transmission which results a net data rate of 109.92 Gb/s. In this case, all the services are modulated with 64QAM mapping formats. The receiver sensitivities are about −21.6 dBm and −17.2 dBm for baseband and radio signals respectively and the performances of the four-wavelength channels are almost the same. We also compare the four-wavelength system performance with single wavelength case. For the baseband signals, it can be seen that the power penalty is about 0.2dB at BER of 10−3 which indicates a good performance if FEC technology is used. Comparing with the one wavelength case, there is about less than 0.1dB power penalty in the WDM case which almost can be ignored. Because in the one wavelength case, the central carrier and couple carriers are spaced at 30GHz and the interference from the couple carriers can be equivalent to that in WDM case. For the radio signals, the power penalty at BER of 10−3 is about 0.4dB after fiber link, which is mainly due to the fiber dispersion during transmission. For the WDM case, the power penalty is a little deteriorated as compared to the single wavelength and can also be ignored. For the radio signal, the main interference comes from the fiber dispersion due to the couple carriers spaced at 60-GHz instead of WDM neighbor channel.
Fig. 5 Measured BER curves of baseband and radio signals with 64QAM constellation mapping and total rate of 109.92 Gb/s.
From Fig. 5, we can see that the system performances of the four-wavelength channels are almost the same, so we only take λ2 (1550.12 nm) for the following analysis. According to the service demand, the speed modes in the leading symbol would be different to each other. Here we demonstrate by setting 16QAM, 32QAM and 64QAM respectively for the three services and all the services adopt the same QAM mapping each time which means the three services averagely share the total access rate. Figure 6 shows the measured BER curves of the three services with access rate of 18.32 Gb/s (12.83 Gb/s for baseband and 5.49 Gb/s for radio), 22.9 Gb/s (16.03 for baseband and 6.87 Gb/s for radio) and 27.48 Gb/s (19.24 Gb/s for baseband and 8.24 Gb/s for radio) respectively. The bandwidth is 3.5 GHz for the baseband signal and 1.5GHz for the radio signal. The receiver sensitivities at BER of 10−3 are −24.3 dBm, −23 dBm and −21.8dBm for baseband signals, and −20.4 dBm, −18.8 dBm and −17.5 dBm for radio signals. Figure 7 shows the constellation diagrams for baseband and radio signals with different access rates respectively. The radio signals seem to suffer more due to the fading effect and fiber dispersion.
Fig. 6 Measured BER curves of the baseband and radio signals with access rate of 18.32 Gb/s, 22.9 Gb/s and 27.48 Gb/s respectively (for λ2 channel).
Besides the above scene, the three kinds of services can take different access rates. There we assume service-1 is 8QAM mapped, service-2 is 16QAM mapped and service-3 is 32QAM mapped respectively. Figure 8 illustrates the measured BER curves when different services are chosen and the cases are shown in Table 1 . The total bandwidth is also 5GHz during measurement. Figure 8 shows the BER performances of baseband and radio signals for the three cases. It can be observed that the receiver sensitivities are −25 dBm, −24.48 dBm and −24dBm for baseband signals and −22.4 dBm, −21.7 dBm and −21.1 dBm for radio signals at BER of 10−3.
Fig. 8 The measured BER curves of the baseband and radio signals with different services and access rates after 30km transmission.
Table 1. System Parameters for Experiment
From Fig. 8, we can see that the receiver sensitivities at BER of 10−3 for baseband and radio 8QAM-OFDM signals are −22.4dBm and −25dBm respectively. The optical path loss is about 13.5dB. For the baseband signal, the margin budget is 16.5dB, which is abundant for a 1:32 split ratio; for the radio signal, the margin budget is 13.9dB, which is close to the split loss of a 1:32 splitter. The maximum number of users for one wavelength can be 64, and the number can be up to 256 in our experiment.
4. Conclusion
We have proposed and experimentally demonstrated a uni-PON architecture for the future all-service access networks. It can offer an infrastructure which enables a hybrid access to support ubiquitous broadband services with different access solutions. Moreover it can provide a dynamic resource allocation and variable access rate by employing a leading symbol at physical-layer which uses adaptive algorithm. In our experiment, a data rate of 109.92-Gb/s with 4 wavelength channels can be reached and evaluated for a 30-km SMF link where each wavelength requires the bandwidth of only 5 GHz. Furthermore, the different services with dynamic resource allocation and variable access rates are also demonstrated in form of baseband and radio signals. The experiment suggests that proposed scheme is a suitable solution for the next generation flexible access networks.
Acknowledgments
The financial supports from National Basic Research Program of China with No. 2010CB328300, National Natural Science Foundation of China with No. 60932004, 61077050, 61077014, 61177085, BUPT Excellent Ph. D. Students Foundation with No.CX201112. The project is also supported by the Fundamental Research Funds for the Central Universities with No.2012RC0311, 2011RC0307.
References and links
1. G. Chang, Z. Jia, J. Yu, A. Chowdhury, T. Wang, and G. Ellinas, “Super-Broadband Optical Wireless Access Technologies,” in Proc. OFC, paper OThD1 (2008).
2. J. Chen, L. Wosinska, C. Machuca, and M. Jaeger, “Cost vs. reliability performance study of fiber access network architectures,” IEEE Commun. Mag. 48(2), 56–65 (2010). [CrossRef]
3. F. Effenberger, “The XG-PON system: cost effective 10Gb/s access,” J. Lightwave Technol. 29(4), 403–409 (2011). [CrossRef]
4. J. Kani, F. Bourgart, A. Cui, A. Rafel, M. Campbell, R. Davey, and S. Rodrigues, “Next-generation PON part I—Technology roadmap and general requirements,” IEEE Commun. Mag. 47(11), 43–49 (2009). [CrossRef]
5. 10-Gigabit-Capable Passive Optical Network (XG-PON) Systems: Definitions, Abbreviations, and Acronyms, ITU-T G.987 (2009).
6. S. Jain, F. Effenberger, A. Szabo, Z. Feng, A. Forcucci, W. Guo, Y. Luo, R. Mapes, Y. Zhang, and V. O’Byrne, “The world’s first XG-PON field trial,” J. Lightwave Technol. 29(4), 524–528 (2011). [CrossRef]
7. D. Kilper, “Energy Efficient Networks,” in Proc. OFC, USA, paper OWI5 (2011).
8. 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]
9. J. L. Wei, A. Hamié, R. P. Gidding, E. Hugues-Salas, X. Zheng, S. Mansoor, and J. M. Tang, “Adaptively modulated optical OFDM modems utilizing RSOAs as intensity modulators in IMDD SMF transmission systems,” Opt. Express 18(8), 8556–8573 (2010). [CrossRef] [PubMed]
10. C. Yeh, C. Chow, and C. Hsu, “40-Gb/s Time-Division-Multiplexed Passive Optical Networks Using Downstream OOK and Upstream OFDM Modulations,” IEEE Photon. Technol. Lett. 22(2), 118–120 (2010). [CrossRef]
11. H. Zhang, G. Pickrell, Z. Morbi, Y. Wang, M. Ho, K. Anselm, and W. Hwang, “32-Channel, Injection-Locked WDM-PON SFP Transceivers for Symmetric 1.25 Gbps Operation,” in Proc. OFC, USA, paper NTuB4 (2011).
12. E. Wong, “Current and Next-Generation Broadband Access Technologies,” in Proc. OFC, USA, paper NMD1 (2011).
13. N. Cvijetic, “OFDM for Next-Generation Optical Access Networks,” J. Lightwave Technol. 30(4), 384–398 (2012). [CrossRef]
14. “Cloud-Radio Access Network (C-RAN) White Paper,” website: http://labs.chinamobile.com/cran/.
15. N. Cvijetic, M.-F. Huang, E. Ip, Y. Shao, Y.-K. Huang, M. Cvijetic, and T. Wang, “1.92 Tb/s coherent DWDM-OFDMA-PON with no high-speed ONU-side electronics over 100 km SSMF and 1:64 passive split,” Opt. Express 19(24), 24540–24545 (2011). [CrossRef] [PubMed]
16. 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]
17. J. Armstrong, “OFDM for Optical Communications,” J. Lightwave Technol. 27(3), 189–204 (2009). [CrossRef]
18. 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]
19. Z. Jia, J. Yu, and G.-K. Chang, “A Full-Duplex Radio-Over-Fiber System Based on Optical Carrier Suppression and Reuse,” IEEE Photon. Technol. Lett. 18(16), 1726–1728 (2006). [CrossRef]
20. P. T. Shih, C. T. Lin, W. J. Jiang, H. S. Huang, J. Chen, A. Ng’oma, M. Sauer, and S. Chi, “Transmission of 20-Gb/s OFDM signals occupying 7-GHz license-free band at 60 GHz using a RoF system employing frequency sextupling optical up-conversion,” Opt. Express 18(12), 12748–12755 (2010). [CrossRef] [PubMed]
21. Y. Hsueh, M. Huang, S. Fan, and G. Chang, “A Novel Lightwave Centralized Bidirectional Hybrid Access Network: Seamless Integration of RoF With WDM-OFDM-PON,” IEEE Photon. Technol. Lett. 23(15), 1085–1087 (2011). [CrossRef]
22. C. H. Yeh, C. W. Chow, H. Y. Chen, and B. W. Chen, “Using adaptive four-band OFDM modulation with 40 Gb/s downstream and 10 Gb/s upstream signals for next generation long-reach PON,” Opt. Express 19(27), 26150–26160 (2011). [CrossRef] [PubMed]
23. J. Ma, J. Yu, C. Yu, X. Xin, and Q. Zhang, “Transmission performance of the optical mm-wave generated by double-sideband intensity-modulation,” Opt. Commun. 280(2), 317–326 (2007). [CrossRef]
