Abstract: We propose and demonstrate a novel optical orthogonal frequency-division multiple access (OFDMA)-based metro-access integrated network with dynamic resource allocation. It consists of a single fiber OFDMA ring and many single fiber OFDMA trees, which transparently integrates metropolitan area networks with optical access networks. The single fiber OFDMA ring connects the core network and the central nodes (CNs), the CNs are on demand reconfigurable and use multiple orthogonal sub-carriers to realize parallel data transmission and dynamic resource allocation, meanwhile, they can also implement flexible power distribution. The remote nodes (RNs) distributed in the user side are connected by the single fiber OFDMA trees with the corresponding CN. The obtained results indicate that our proposed metro-access integrated network is feasible and the power distribution is agile.
©2013 Optical Society of America
Driven by increasing demand and new service requirement, the access bandwidth is upgraded in several orders of magnitude, and as both the access capacity and the number of remote end-users continue to rise, the various types of traffic need to be aggregated and exchanged in the metropolitan area network (MAN), thus the access capacity of MAN will soon be exhausted because of its limited throughput. Furthermore, the separated metro and access networks resulting in a complex infrastructure and consuming a very large number of network elements, which make the operator’s capital expenditure (CapEx) and operating expense (OpEx) very expensive [1, 2]. Consequently, more attention has been drawn to the integration of metro and access networks, the long-reach access directly connected to the backbone networks is regarded as the best candidate to fulfill this purpose, and hence the long-reach passive optical network (LR-PON) emerges as the times require . Additionally, the optical orthogonal frequency-division multiple access-based passive optical network (OFDMA-PON) has been widely studied due to its high spectrum efficiency and flexible access [4, 5].
The LR-PON commonly adopts an aggregation network, a local exchange or a reach extender between the optical network unit (ONU) and the optical line terminal (OLT) to obtain long distance, high capacity and high split-ratio [3, 6]. The LR-PON, nowadays, usually exhibits a ring-and-spur topology instead of the traditional tree-and-branch topology of PON. Meanwhile, the access segment is a tree-and-branch architecture . Thus, the LR-PON now is characterized by simple scalability and upgradeability, multi-operability and high bandwidth for each user. There exist two typical metro-access integrated network architectures, one is the FP7 European research project “Scalable Advanced Ring-based passive Dense Access Network Architecture” (SARDANA) which merges a WDM double-fiber ring with TDM single fiber access trees by means of passive add/drop remote nodes (RNs) [8, 9]. Another is the Metro Access Ring Integrated Network (MARIN) which forms a mesh network through interconnecting multiple access ring networks. The MARIN employs dense WDM technology to achieve capacity enhancement and traffic routing, dynamic wavelength allocation was used to provide resource sharing at the same time. Each tree network occupies specific wavelength channels and a cyclic array wavelength grating was used for wavelength routing [9, 10]. Metro-access integrated network with all-optical virtual private network function and converged passive metro-access network with remotely pumped amplifiers have been also reported . In the above mentioned schemes, they all employ the mix of WDM and TDM to obtain sustained data rate with highly shared infrastructure. Nevertheless, the TDM access has a strict requirement for time synchronization for an acceptable performance, with the purpose of supporting vast end users, the WDM access needs equipped with a mass of costly tunable lasers . Obviously, these access approaches are either too complicated or not cost-effective enough. In contrast, optical OFDMA technology is a more attractive candidate for high speed metro-access integrated network due to its high spectrum efficiency, robust dispersion tolerance and flexible access featuring a dynamic bandwidth allocation and a multi-rate adjustment [13, 14].
In this work, an optical OFDMA single fiber ring and high-quality OFDMA access tree enabled metro-access integrated network is proposed and successfully verified. It consists of an OFDMA uni-bidirectional single fiber ring and multiple OFDMA bidirectional single fiber access trees, and they are interconnected by optical central nodes (CNs) which implement the traffic routing. As new CNs can be installed according to the need of the users, the integrated network offers enhanced scalability, flexible distribution and very cost-effective.
2. Principle of the proposed optical metro-access integrated network
Figure 1(i) depicts the principle of our proposed optical OFDMA-based metro-access integrated network. It consists of an OFDMA single fiber ring in which n CNs are distributed and multiple OFDMA single fiber trees, these trees access the remote nodes (RNs) in the user terminal side, and a unique CO connects the core networks. The CO responsible for data aggregation, resource allocation and the control function of the entire network. The traffic routing is performed by the CNs, and the CNs can be added according to the need of the users and they transmit/receive information independently, each CN can serve multiple RNs through the splitter in the access segment. Users are located at the RNs. The transmission of information is fulfilled by the sub-carrier allocation of OFDM. First, the CO transforms the signal from the backbone networks into OFDM signal as shown in the Fig. 1(i) marked with blue rectangle blocks, and the OFDM signal is divided into several orthogonal sub-bands served for a mass of RNs, each sub-band contains multiple subcarriers. All the sub-bands are transmitted to each RN while the RNs just demodulate the corresponding sub-band, namely, the RN can choose either one of the sub-band or various sub-bands in accordance with the need of users. Then, the downstream OFDM signal is sent into the first CN, namely CN-1, the CN delivers the signal combined with the upstream optical carrier to the RNs connected to it afterwards and receives the information from the corresponding RNs, here we assume that the RNs just upload one sub-band. Next, CN-1 launches the downstream OFDM signal together with the upstream signal into next CN, namely CN-2, the frequency domain description of the information transmission is shown in Fig. 1 along the ring, and the orange rectangle blocks represent the upstream signal, we note that each CN must carry all the sub-bands not a part of them. Similarly, CN-2 delivers the signal to CN-3, in this way, CN-n-1 sends the OFDM signal to CN-n, and all the signals are converged in the CO finally.
Figure 1(ii) shows the power distribution of the proposed OFDMA-based metro-access integrated network. We can see that the total power is P, they are divided into two parts when comes to the CN-1. One part Pg1α1 is assigned to the RNs connected to the CN-1, where α1 is the power distribution factor of CN-1, and g1 represents the loss and gain factor of CN-1 which can be written to g1 = γ1 + δ1, where γ1 is the loss factor and δ1 is the gain factor of CN-1, respectively. Another part Pg1(1-α1) is distributed to the next CN-2. Consequently the power is divided into two parts at CN-i. PitoRN is the signal sent in to RN, while PitoCN is the signal transmitted in to the next CN, and they can be expressed as follows,
3. Experimental setup, results and discussion
Figure 2 illustrates the proof-of-concept experimental setup for the proposed OFDMA-based metro-access integrated network. The ring covered 65 km range, at 10 Gb/s down and up stream, with 3 CNs. The subcarriers are divided into three bands: band 1, band 2 and band 3. Each band carries different service traffic, they are distributed dynamically according to the need of users.
In the CO, the continuous wave (CW) lightwave is generated by a distributed feed back (DFB) laser at 193.1 THz with 10 MHz linewidth and 2.0 mW launch power. A Mach-Zehnder modulator (MZM) of 30-dB extinction ratio (ER) is driven by the 7.5-GHz radio-frequency (RF) DS 16 quadrature-amplitude modulation (QAM) OFDM signal produced by the OFDM transmitter (Tx), as shown in the inset of Fig. 2, to generate the optical OFDM signal. In the OFDM-Tx, a 10-Gb/s baseband 16 QAM-OFDM signal is generated off-line with the 128 FFT size, while CP of 1/5 is used. Then the baseband OFDM signal is up-converted to 7.5 GHz RF using an analog inphase/quadrature (IQ) mixer and a 7.5 GHz RF source. The modulated signal is converted to single-sideband (SSB) signal format through a tunable optical filter (OF) with 18 GHz bandwidth, and then the optical signal is launched into the standard single mode fiber (SSMF) with the attenuation of 0.2 dB/km and the dispersion of 20 ps/nm·km. After 20 km fiber, the SSB-OFDM signal comes to the first CN, namely the CN-1, and likewise, the signal coming out from the CN-1 arrives at the CN-2 after another 30 km fiber. Take the CN-2 for example, the signal is firstly amplified by an Erbium-doped fiber amplifier (EDFA), and then it is split into two parts in the CN by an optical splitter (OS). Dynamic power allocation can be achieved by setting various couple-factors, and the couple-factor is different in each CN, we can find a balance between the gain of the amplifier and the couple-factor to obtain a better performance. And then an OF is used to extract the downstream signal sub-bands from one part of the splitting signal according to the need of RNs. Next, the signal is combined with US seeding laser centered at 193.132 THz through an OC. After that, the combined signal is launched into a 5 km SSMF for the RNs. The RN-2 is one of the user terminals in the OFDMA access tree, where an OF is utilized to separate the DS-OFDM signal and the US optical carrier. The DS-OFDM signal is detected by a photodiode (PD) with the responsivity of 1.0 A/W, and then received by the OFDM receiver (Rx) which is illustrated in the inset of Fig. 2. In the OFDM-Rx, a 7.5 GHz RF source is utilized to down-convert the RF-OFDM signal in an analog IQ-mixer. After down-conversion and analog to digital conversion (ADC), the produced DS baseband 16 QAM-OFDM signal is analyzed off-line to get the DS bit-error-rate (BER) performance. The US optical carrier is modulated through a MZM which is driven by a 10 Gb/s US-16QAM-OFDM signal produced via the OFDM-Tx. The modulated US-OFDM signal is converted to SSB signal format through an OF, and then transmitted back to the CN-2 through the 5 km SSMF. The other part of the splitting signal is delivered to CN-3 combined with the US-OFDM signal through 10 km fiber. Finally, all the DS-OFDM signals are distributed to the corresponding end users, and all the US-OFDM signals are converged in the CO.
Figure 3(i) –3(vi) show the results of the spectra distribution of the corresponding points in Fig. 2. Figure 3(i) presents the electrical spectrum of the RF DS-OFDM signal which is divided into three bands for different RNs. Figure 3(ii) gives the double-sideband (DSB) DS OFDM signal. The filtered SSB signal is illustrated in the Fig. 3(iii), and we can clearly find that the divided three bands are b1, b2 and b3. The signal sent into CN-2, as given in Fig. 3(iv). It can be seen that the US-OFDM signal b1' from the RN connected to the CN-1 is added. The signal send into the RN-2 which is connected to the CN-2, as shown in Fig. 3(v), the downstream signals in CN-2 and the US optical carrier centered at 193.132 THz are clearly found in Fig. 3(v). Even though all the sub-bands are transmitted to RN-2, the OFDM-Rx can only demodulate the sub-bands it needed, and the dynamical subcarrier allocation can be achieved by this way. The signal launched into the CN-3 is presented in Fig. 3(vi), the US-OFDM signals b1', b2' at 193.132 THz coming from RN-2 are successfully added.
Figure 4 shows the BER performances versus the received optical power and the corresponding constellations with three RNs for two cases, one case is the two coupler factors are 0.9 and 0.6 respectively, other case is the two coupler factors are 0.8 and 0.6 separately. We can find that the performance of three RNs is relative average in the first case, while the performance exist big difference between RN-1 and the two RNs latter for case 2. From Fig. 4 we can draw a conclusion that as the transmission distance increased, the BER performance has some deterioration and the constellations get radiation. Meanwhile, the constellations have rotation to some degree because of the phase noise. In general, the performance of various RNs for different cases can be acceptable.
Figure 5(i) plots the BER performance versus the received optical power and the corresponding constellations for uplink of our proposed metro-access integrated system after 15-km, 45-km and 65-km SSMF transmission, respectively. A received optical power of about −18.7 dBm is needed to reach the forward-error-correction (FEC) limit of BER = 10−3. From the comparison of the 15-km, 45-km and 65-km SSMF transmission, we can see that the power penalty induced by the fiber dispersion is almost negligible. Moreover, it is obvious that the 16-QAM constellations are effectively recovered. Figure 5(ii) illustrates the contrast of the received optical power for two cases when the BER comes to 10−3. The horizontal ordinate represents the RN, here are RN-1, RN-2 and RN-3. We can clearly find that the FEC limit of three RNs is around −19.4 dBm for case 1, and the power differentials between RN-1 and RN-3 is less than 1 dBm. However, the performance difference among three RNs become larger for case 2 even though the FEC limit of RN-1 comes to −20.9 dBm, because the power differentials between RN-1 and the two RNs latter is about 1.8 dBm. On the whole, the performance of case 1 is relatively better than case 2 and the FEC limit of all RNs for both cases is acceptable.
We have proposed and successfully demonstrated a novel cost-effective OFDMA based metro-access integrated network with dynamic resource allocation in this work. In our proposed network, we employ the reconfigurable OFDMA single fiber ring and tree topology to fulfill the integration of two separate networks and achieve high scalability and flexibility. Moreover, the subcarriers can be allocated dynamically to different RNs according to the users’ requirements and the power can distribute dynamically to achieve a better performance. The results show that the proposed scheme is feasible and may be viewed as a highly attractive candidate for next-generation broadband network.
This work is jointly supported by national natural science foundation of China (No. 61171045), and open fund of national engineering center of HUST (No. NK201202).
References and links
1. J. M. Finochietto, J. Aracil, Á. Ferreiro, J. P. Fernandez-Palacios Gimenez, and O. Gonzalez de Dios, “Migration strategies toward all optical metropolitan access rings,” J. Lightwave Technol. 25(8), 1918–1930 (2007). [CrossRef]
2. N. Cvijetic, M. Cvijetic, M. F. Huang, E. Ip, Y.-K. Huang, and T. Wang, “Terabit optical access networks based on WDM-OFDMA-PON,” J. Lightwave Technol. 30(4), 493–503 (2012). [CrossRef]
3. H. Song, B. W. Kim, and B. Mukherjee, “Multi-thread polling: a dynamic bandwidth distribution scheme in long-reach PON,” IEEE J. Sel. Areas Comm. 27(2), 134–142 (2009). [CrossRef]
5. C. F. Zhang, C. Chen, J. Huang, and K. Qiu, “Performance improvement of optical OFDMA-PON using data clipping and additional phase,” IEEE Photon. Technol. Lett. 24(4), 255–257 (2012). [CrossRef]
6. B. Skubic, J. Chen, J. Ahmed, B. Chen, L. Wosinska, and B. Mukherjee, “Dynamic bandwidth allocation for long-reach PON: Overcoming performance degradation,” IEEE Commun. Mag. 48(11), 100–108 (2010). [CrossRef]
7. E. Wong, “Next-generation broadband access network and technologies,” J. Lightwave Technol. 30(4), 597–608 (2012). [CrossRef]
8. J. Prat, J. Lazaro, P. Chanclou, R. Soila, A. M. Gallardo, A. Teixeira, G. M. TosiBeleffi, and I. Tomkos, “Results from EU project SARDANA on 10G extended reach WDM PONs,” in OFC/NFOEC, Paper OThG5 (2010).
9. T. Pfeiffer, “Converged heterogeneous optical metro-access networks,” in ECOC, Paper Tu.5.B.1 (2010).
10. S. Wong, W. Shaw, N. Cheng, C. M. Qiao, and L. Kazovsky, “Dynamic wavelength allocation in a converged and scalable interface for metro-access ring integrated networks,” in OFC/NFOEC, Paper OTuI6 (2008).
11. Y. Tian, L. Leng, and Y. K. Su, “A metro-access integrated network with all-optical virtual private network function using DPSK/ASK modulation format,” in SPIE, Paper 71370O (2008).
12. D. Z. Hsu, C. C. Wei, H. Y. Chen, W. Y. Li, and J. Chen, “Cost-effective 33-Gbps intensity modulation direct detection multi-band OFDM LR-PON system employing a 10-GHz-based transceiver,” Opt. Express 19(18), 17546–17556 (2011). [CrossRef] [PubMed]
14. C. Chen, C. F. Zhang, D. M. Liu, K. Qiu, and S. Liu, “Tunable optical frequency comb enabled scalable and cost-effective multiuser orthogonal frequency-division multiple access passive optical network with source-free optical network units,” Opt. Lett. 37(19), 3954–3956 (2012). [CrossRef] [PubMed]