Open Access
Add to BibSonomyAbstract
Optical multicasting that supports point-to-multipoint traffic replication can be one of the necessary techniques in next-generation all-optical elastic networks. In this paper, we propose an optical multicasting approach for polarization-division-multiplexing (PDM) orthogonal frequency division multiplexing (OFDM) signals based on a novel polarization-interleaved multi-pump (PIMP) four-wave mixing (FWM) scheme in highly nonlinear fiber (HNLF). Besides format transparency and the support of PDM signals, the scheme further enables wide spectral tunability of generated replicas. The pump frequency arrangement for the scheme is presented, which successfully prevents the replicas from being superimposed by unwanted FWM components during tuning. We experimentally demonstrate multicasting operation of a 3-band 100-Gb/s PDM-OFDM signal. With different input positions, 1.4 and 1.6 Terahertz tuning ranges of four replicas are achieved with Q-factor performance better than the forward error correction threshold. Tunable replica spacing from 100-GHz to 250-GHz are also verified. In addition, the scalability of the scheme is demonstrated via 5-pump multicasting, successfully generating a total of 14 replicas.
© 2016 Optical Society of America
1. Introduction
Today’s traffic demand in backbone networks is becoming increasingly bandwidth-hungry and diversified. To address the requirements of higher throughput, higher energy efficiency and lower latency, it is necessary to develop next-generation all-optical and elastic networks [1–4]. Such networks would exploit spectrally efficient signal types like polarization-division-multiplexing (PDM) orthogonal frequency division multiplexing (OFDM) [5] and Nyquist [6] signals with amplitude and phase modulation, as well as all-optical processing techniques capable of manipulating these advanced signals. All-optical functions for elastic networks, such as reconfigurable optical add-drop multiplexing (ROADM), sub-band switching, distributed superchannel aggregation and frequency conversion, have been investigated [7–12].
Besides, optical multicasting of optical signals can be another technique in demand, whereby a traffic is simultaneously replicated multiple folds in a single optical-layer device, satisfying the requirement of point-to-multipoint communication in the network, e. g., inter-datacenter service migration and data back-up [33]. Compared to state-of-the-art multicasting implemented in IP layer, multicasting directly in optical domain can circumvent the electronic bottleneck and can support much higher data rate such as 100-Gb/s and beyond. In this work, we focus on optical multicasting in spectrum domain. Compared to broadcasting based on power split, spectral multicasting relieves the spectrum continuity constraint in spectrum allocation process, so that link spectrum usage can be enhanced and traffic blocking due to spectrum contention may be reduced [13–16]. In future all-optical elastic networks, the following issues should be addressed in optical multicasting:
- (1) Independence of modulation format, i. e., amplitude and phase transparency, since traffic in the network can be mixed-format.
- (2) Supporting of polarization division multiplexing (PDM) signals, as PDM would be extensively utilized to double spectrum efficiency.
- (3) Flexibility of input/output spectral positions, so as to fully facilitate the advantage of spectrum allocation. In particular, the spectral tunability of generated replicas is highly desirable for routing and spectrum allocation (RSA) of multicasting connections. In fixed-grid WDM optical networks, the benefit of wavelength conversion on optical multicasting has been presented in terms of link resource utilization [15]. Similarly in flex-grid networks, with greater tunability, the replicas can be freely distributed to available spectrum on the downstream links, thus allowing more flexible management of spectrum resource.
Previously, efforts have been made on all-optical wavelength multicasting [17–26]. While most works only focus on multicasting with fixed conversion range, several have explored spectrally tunable multicasting schemes. For example, multicasting of non-return-to-zero (NRZ) signal using fiber optic parametric amplifiers is demonstrated with ~64 nm tuning range [23]. However, limited by the principle (pump-modulated four-wave mixing), this scheme is not format-transparent. In [24], the input signal is firstly broadened to a super-continuum by self-phase modulation (SPM) and then tunably filtered to create multiple tunable replicas, which also doesn’t possess format transparency. More recently, a QAM multicasting scheme in periodically-poled Lithium Niobate (PPLN) is proposed [25], in which each replica is in principle tunable by tuning its corresponding differential frequency generation (DFG) pump laser. However, the damage power threshold of PPLN may limit its scalability towards more replicas. In an alternative PPLN-based scheme [26], a 3-tone generator based on MZM is exploited to produce 5 replicas using only a single pump laser. But the tuning range of replicas is only ~25-GHz, which is strongly limited by the RF frequency imposed on the MZM and the quasi phase matching (QPM) bandwidth of PPLN.
By far, tunable and transparent multicasting of optical PDM-OFDM signals has rarely been experimentally investigated. In this paper, we propose and experimentally demonstrate an optical multicasting approach for multi-band PDM-OFDM signals based on FWM in HNLF. Different from co-polarized multi-pump scheme in which the tunability is limited to ~300-GHz [27], a novel scheme exploiting polarization-interleaved multi-pumps (PIMP) is proposed to generate signal replicas with much wider tunability. In the scheme, the replicas generated by orthogonal pump pairs (termed as RO) can be widely tuned by shifting all pumps together. A pump frequency arrangement is presented, which prevents the replicas from being interfered by unwanted FWM components and ensures signal integrity during tuning. At different input positions, the experiments demonstrate tuning ranges of 1.4~1.6 Terahertz in C-band for 4 replicas. Meanwhile, the spacing between replicas can also be tuned by changing pump spacings. Moreover, additional replicas can be generated by co-polarized pump pairs (termed as RC), which can be another option of replica allocation for multicasting.
The remaining of this paper is organized as follows. The technique principle and scalability of the proposed PIMP scheme is discussed in Section 2. The experiment setup is described in Section 3. In Section 4 we study in detail PIMP-based 4-pump tunable multicasting of a 3-band, 100-Gb/s PDM-OFDM signal. To verify the scalability of the scheme, 5-pump PIMP-based multicasting that concurrently generates 14 PDM-OFDM replicas (6 RO and 8 RC) is also demonstrated. Finally, Section 5 concludes the paper.
2. Operation principle of PIMP-based tunable optical multicasting
The conceptual spectrum of PIMP-based multicasting is illustrated in Fig. 1, where in Fig. 1(c) only pumps are shown. Assume that the original signal centered at fS with a bandwidth of BS needs to be multicast. A total of (MH + MV) polarization interleaved pumps are placed at {fHj, j = 1,2,…,MH, fVk, k = 1,2,…,MV}, in which MH pumps (H-polarization, red) are co-polarized while being orthogonal to the other MV pumps (V-polarization, blue). Note that the “H” or “V” polarization of pumps is independent of the “X” or “Y” polarization of the PDM-OFDM signal. In the process of FWM, every two orthogonal pumps {fHj, fVk} interact with the original signal and create one RO centered at fHj + fVk−fS [30]. Therefore, a total of (MH × MV) RO-type replicas can be produced in principle. However, multiple undesired FWM components are also created. Desired replicas and undesired FWM components may be superimposed if frequency relationships of FWM are not carefully configured, in which the disruptive interference would induce substantial demodulation errors at the receiver. To guarantee the signal integrity of all replicas, here we propose an effective solution for pump frequency arrangement which ensures that: (a) pump idlers don’t fall on any replicas; (b) replicas don’t overlap each other; (c) degenerate FWM components centered at {2fHj−fS} and {2fVk−fS} don’t fall on any replicas. In the scheme, the spacing between any two pumps is an integer multiple of Δf = BS + BG, in which BG represents a guard-band to accommodate pump idlers. Meanwhile, the frequency spacing between pumps and original signal follows:
fM and f1 in Eq. (1) respectively denotes the frequency of the pump with smallest and largest frequency spacing to fS. Larger p corresponds to less polarization sensitivity of the multicasting operation [30]. p is kept larger than 8 in the experiments.
Fig. 1 (a)-(c) Conceptual spectrum of PIMP-based 3-, 4-, and 5-pump multicasting. Black blocks represent original signal. Red/blue arrows represent H-pol./V-pol. pumps. Yellow blocks represent replicas. Blue blocks represent degenerate FWM components. (d) Proposed structure of multicasting module. PM-WSS: polarization-maintaining wavelength selective switch. BV-WSS: bandwidth-variable wavelength selective switch.
Since each RO is centered at fHj + fVk−fS and degenerate FWM components are centered at 2fHj−fS or 2fVk−fS, the pump frequencies are arranged such that conditions (b) and (c) are satisfied:
Figure 1(a) gives 3-pump solution to generate 2 RO. As shown in Figs. 1(b) and 1(c), the pump frequency solutions for generation of more RO is obtained by recursively placing each pump orthogonal to the prior-placed pump considering Eq. (2)-(3). As such, PIMP-based multicasting generating 4 and 6 RO-type replicas can be implemented using 4 and 5 pumps. The replica positions can be tuned by increasing/decreasing p in Eq. (1) given that FWM efficiencies (and OSNRs) of replicas are guaranteed. Meanwhile, the replica spacing can be tuned by tuning Δf in Eq. (1). Therefore two degrees of freedom are offered for tuning RO.Meanwhile, each two co-polarized pumps (e. g., {fHj, fHk}) also mix with the original signal and generate 2 replicas centered at fS + (fHj−fHk) and fS−(fHj−fHk) [28, 29]. Therefore, MH × (MH−1) + MV × (MV−1) additional RC-type replicas can be generated given that RC don’t overlap, or equivalently, Eq. (4) is satisfied in addition to Eq. (2)-(3).
Note that if Eq. (4) is not fulfilled, some RC-type replicas would overlap and the amount of usable RC-type replicas would decrease, but the amount of usable RO-type replicas would not be affected. Pump frequency arrangements in Figs. 1(a)-1(c) have taken Eq. (4) into account. Moreover, the spacing between a RC and the original signal can be tuned by tuning Δf in Eq. (1). Finally, we note that RO are spectrally reversed while RC are spectrally non-reversed with respect to the original signal.We also present a possible structure of multicasting subsystem for the PIMP-based multicasting, as shown in Fig. 1(d). A polarization beam combiner (PBC) combines outputs from H-pol. pump pool {fHj} and V-pol. pump pool {fVk}. Then the pumps are coupled with the signal and launched into the HNLF. After multicasting, a multi-port bandwidth-variable wavelength selective switch (BV-WSS) is controlled to select and forward the desired replicas. The unwanted FWM components can also be filtered off. Moreover, we propose here a pump pool structure composed of an optical comb followed by a 1 × 2 polarization-maintaining wavelength selective switch (PM-WSS) (see the inset). The optical comb simultaneously provides two pump pools while the WSS reconfigurably selects comb lines as pumps. Compared to laser banks, such implementation may have better scalability and avoids FWM phase noise transfer for RC-type replicas [27, 35, 36].
3. Experimental setup
The experiment setup of PIMP-based tunable multicasting is depicted in Fig. 2. 4 pumps are used to generate 4 RO and 4 RC of a 3-band PDM-OFDM signal. In the PDM-OFDM transmitter (Tx), the coupled output of three 12.5-GHz spaced lasers is fed into an IQ modulator and modulated by 12.5-GS/s quadrature phase shift keying (QPSK) DFT-spread OFDM signal. In each DFT-spread OFDM symbol, 204-point FFT, 256-point IFFT, 4 pilots and 12-sample cyclic prefix (CP) are adopted. PDM is emulated with a polarization controller (PC), a polarization beam splitter (PBS), an optical delay line with delay of exactly one DFT-spread OFDM symbol period, and a polarization beam combiner (PBC). Considering the overhead of 7% for FEC, 4% for training symbols, 4.48% for CP and 1.49% for pilots, the 3-band PDM-OFDM signal has a gross data rate of 119.4-Gb/s and a net data rate of 100.5-Gb/s. In the multicasting subsystem, 2 coupled pumps (“H-pol. pumps”) and another 2 coupled pumps (“V-pol. pumps”) are combined by a polarization beam combiner. The linewidth of each pump laser is about 100-kHz. After passing through a polarization controller, the pumps are coupled with the signal. They are launched into the HNLF for multicasting. Our available HNLF has a length of 1-km, a nonlinear coefficient of ~11 W−1/km, an attenuation coefficient of ~0.9 dB/km, dispersion slope of ~0.02 ps/(nm2•km) at wavelength region of 1540~1560-nm and zero dispersion wavelength of around 1551-nm. After multicasting, a multi-passband optical filter (Finisar Waveshaper, 50-GHz per passband) is firstly used to filter out all replicas before amplification. Then a BV-WSS (Finisar Waveshaper) selects and forwards each replica using 50-GHz passband. Note that the first filter is used just to satisfy the input power range of our available EDFA. In principle, a single WSS (with EDFA) is sufficient to select replica(s). When bit-error rate (BER) versus optical signal-to-noise ratio (OSNR) is measured, OSNR is adjusted by varying the loaded ASE noise. Finally, a digital coherent receiver (Rx) detects the replicas. The linewidth of the local oscillator (LO) is about 100-kHz. The offline digital signal processing procedure mainly includes digital low-pass filtering, frame and frequency synchronization, training-aided channel estimation, 2 × 2 multiple input multiple output (MIMO) channel equalization, and pilot-aided phase estimation.
4. Experimental results
4.1 Investigation of spectral tunability of replicas
In order to validate the tunability of replicas in PIMP-based multicasting, the following experimental scenarios are conducted and investigated.
Scenario A: the input signal centered at 191.775-THz (1563.25-nm) is multicast to the large-frequency-side of C-band. Here Δf = 50-GHz. First, we fix the pump and replica frequency, and optimize signal and pump power by measuring the Q-factor performance (derived from constellation expansion [7]) of RO1. In this measurement, “H-pol.” pump lasers are placed at 193.25-THz and 193.05-THz, while “V-pol.” pumps are at 193.1-THz and 193.0-THz. The corresponding RO1~RO4 are at 194.575~194.275-THz with 100-GHz spacing. Figures 3(a) and 3(b) depict spectra at the input of HNLF and output of HNLF. Figure 3(c) shows spectra of 4 replicas at the output of the BV-WSS. We can see that unwanted FWM components have been filtered off. Figure 4(a) shows Q-factor versus signal power (measured after passing OBPF) when the total pump power is fixed at 23-dBm. The optimal signal power is found to be 11.25-dBm. Lower signal power results in insufficient conversion efficiency while higher signal power causes more undesired nonlinearity such as cross-phase modulation. Figure 4(b) shows Q-factor versus total pump power when the signal power is fixed at 11.25-dBm. It is observed that Q-factor is improved as pump power increases from 16-dBm to 23-dBm, which is the maximum output power of our available EDFA. Increasing further pump power (by employing EDFA with larger output power) may improve the signal quality of replicas, but there will be an optimum pump power that corresponds to a high limit of signal quality. Excessive pump power will intensify the Stimulated Brillouin scattering (SBS) effects and degrade the signal quality. In the following experiments, the signal power and pump power are set to 11.25-dBm and 23-dBm, respectively.
Fig. 4 (a) Q-factor of RO1 vs. signal power (pump power = 23 dBm). (b) Q-factor of RO1 vs. pump power (signal power = 11.25 dBm)
Then the center frequencies the replicas are tuned by tuning 4 pumps together (with fixed pump spacing) while the original signal is fixed at 191.775-THz (1563.25-nm). We tune RO1~RO4 from 194.575~194.275-THz (around 1542-nm) to 192.975~192.675-THz (around 1555-nm). According to the principle of FWM, a K-GHz tuning of pumps corresponds to a 2K-GHz tuning of replicas (RO1~RO4). Figure 5(a) plots such linear relationship between pump frequencies and replica frequencies, in which “the largest frequency of pumps” is used as vertical axis of figure and represents spectral positions of all pumps. The measured Q-factor results of RO1~RO4 are shown in Fig. 5(b). As reference, the measured Q-factor of the signal in back-to-back case is 17-dB. All replicas after frequency tuning exhibit Q-factor performance above the 7% FEC limit of Q = 8.53-dB. The constellations of original signal and the replica with lowest Q-factor performance during tuning (RO2 at 194.475-THz) are shown in Figs. 5(c) and 5(d). The polarization state of the PC on the pump branch is not intentionally adjusted, and both polarizations of PDM replicas always have similar performance.
Fig. 5 Scenario A: (a) Frequency relationship between pumps and replicas. (b) Q-factor vs. replica frequency. (c) constellations of original signal. (d) constellations of RO2 at 194.475-THz. (e) OSNR-BER curve of 4 replicas.
During tuning, it is also observed that the Q-factors of the replicas quickly decrease when the replicas are tuned to frequencies larger than 194.575-THz. This is mainly due to insufficient FWM conversion efficiencies, which depend much on the phase mismatch and the nonlinear coefficient of the available HNLF [29, 31]. If larger tuning range is desired, we can employ HNLF with larger nonlinear coefficient, lower dispersion and lower dispersion slope in the scheme [31, 32, 34]. Filtering out-of-band noise on the pump branch [32] may also be helpful.
In this scenario, we also investigate the OSNR-BER performance of replicas. The replicas are at 194.575~275-THz with a spacing of 100-GHz, the spectra of which have been presented in Fig. 4(a). The measured results are shown in Fig. 5(e). Compared to back-to-back case, less than 1-dB OSNR penalties of replicas are observed at 7% FEC limit of BER = 3.8e-3, which are mainly due to the nonlinear crosstalk from adjacent degenerate FWM components and pump-to-signal phase noise transfer [35, 36].
Scenario B: the input signal centered at 194.275-THz (1543.13-nm) is multicast to the small-frequency-side of C-band. Here Δf = 50-GHz. At the beginning of the tuning, the frequencies of the pumps are: fH1 = 193.8-THz, fH2 = 193.7-THz, fV1 = 193.75-THz, fV2 = 193.55-THz. So RO1~RO4 are centered at 193.275~192.975-THz. We then tune RO1~RO4 from 193.275~192.975-THz (around 1552-nm) to 191.875~191.575-THz (near 1565-nm), by tuning 4 pumps together. Similar with Fig. 5(a), the relationship between pump frequencies and replica frequencies is depicted in Fig. 6(a). The measured Q-factor results of RO1~RO4 after each tuning step are shown in Fig. 6(b). All Q-factor performances are also well above 7% FEC limit.
Fig. 6 Scenario B: (a) Frequency relationship between pumps and replicas. (b) Q-factor vs. replica center frequency.
To conclude, in this subsection we show PIMP-based 4-pump multicasting of a 3-band optical PDM-OFDM signal. The frequency spacing between 4 replicas and the original signal can be more than 2.5-THz. At different input positions, replica tuning ranges of 1.4~1.6-THz in C-band are demonstrated with available HNLF.
4.2 Tunability of replica spacing
The replicas in the PIMP-based multicasting can be tuned not only by changing p in Eq. (2) but also by tuning Δf in Eq. (1). In this sub-section we demonstrate the tunability of frequency spacing between replicas by tuning Δf. Moreover, we also evaluate the performance of RC-type replicas in PIMP-based multicasting scheme.
Scenario C: The original PDM-OFDM signal is centered at 194.275-THz (1543.13-nm). Two H-pol. pumps at and two V-pol. pumps are used for multicasting. 4 RO and 4 RC are generated. The frequency spacing between adjacent replicas is tuned to 100-GHz, 200-GHz and 250-GHz by changing Δf in Eq. (1) to 50-GHz, 100-GHz and 125-GHz. The corresponding spectra of HNLF output in Figs. 7(a)-7(c), where the input signal, pump, RC and RO are distinguished by different colors. Figures 8(a)-8(f) show measured Q-factors of newly-generated replicas in these cases. We use “relative frequency of replicas” as horizontal axis of the figures to illustrate the tuning of replica spacing more clearly. The zero frequency in the figures represents the center frequency of 4 RC or that of 4 RO. Q-factors of all replicas including 4 RO and 4 RC in all measured cases are all above the FEC limit. Provided that the pumps are tunable, Δf could be flexibly tuned to other values to satisfy different requirements of spectrum utilization. It is observed that Q-factor of RC1 decreases as the signal-replica spacing increases to 400-GHz (replica spacing of 200-GHz) or more, whereas no obvious performance degradation occurs among RO. The reason is that the tuning range of RC spacing is limited by the conversion range of co-polarized pump FWM, while that of RO spacing can be larger owing to the principle of orthogonal pump FWM [28, 30]. When the spacing of orthogonal pumps and that of co-polarized pumps are significantly smaller than pump-signal spacing, RO would have lower conversion efficiencies than RC [28, 30]. To improve conversion efficiencies and OSNR of RO, one can increase pump power while suppressing SBS effects and/or increasing SBS threshold [19].
4.3 Scalability of the multicasting scheme
As the number of network subscribers continues to grow, it is desired to achieve more signal replicas in certain multicasting scenarios. Here we also investigate the scalability of PIMP-based signal multicasting towards more replicas.
Scenario D: 5-pump multicasting of PDM-OFDM signal is considered, which simultaneously generates 8 RC and 6 RO. The parameters of the signal are the same as those in Scenario B (3-band, 100-Gb/s, centered at 194.275-THz). An additional H-pol. pump is included in the experiment setup. Based on the principle of PIMP scheme, frequencies of H-pol. pumps are configured to be 193.25-THz, 193.15-THz, 192.85-THz, while those of V-pol. pumps are 193.2-THz and 193.0-THz. So the 6 RO are generated at 192.175, 192.075, 191.975, 191.875, 191.775, and 191.575-THz. The 8 RC are generated at 194.675~194.375-THz with spacing of 100-GHz and 194.175~193.875-THz with spacing of 100-GHz. The input signal power and pump power are 11.25-dBm and 23-dBm. The spectrum at the output of HNLF is shown in Fig. 9(a). The 8 RC and 6 RO are marked in the figure. We measure the signal quality of all generated replicas, as shown in Fig. 9(b). The original signal after having traversing the multicasting subsystem is also measured. All 14 replicas exhibit Q-factor above the 7% FEC limit. Compared to original signal, the limiting factors of replica performance are decreased OSNRs and nonlinear crosstalk from adjacent components. The performance fluctuation among replicas is partly because that some replicas (e. g., RO4) suffer relatively strong nonlinear crosstalk induced by the two surrounded strong FWM components (as can be seen from the spectrum). These stronger components are in fact the superimposition of multiple degenerate FWM products. Therefore, one solution to mitigate nonlinear crosstalk is to avoid such superimposition by modifying pump frequency arrangement in PIMP scheme. In addition, when the multicasting scheme is employed to offer higher replica count, signal power and the power of each pump should be further optimized in order to guarantee signal quality of all replicas.
Fig. 9 Scenario D (5-pump PIMP-based multicasting): (a) spectra of HNLF output. (b) Q-factor of all replicas and original signal.
5. Discussion and conclusion
Aside from the proposed all-optical approach, tunable multicasting based on multiple optical-electrical-optical (O/E/O) conversions is also feasible for PDM-OFDM signals, which would provide greater flexibility for signal copies. However, all-optical scheme that bypasses all O/E/O modules can be advantageous in processing speed, cost and energy consumption. To address the limitation of non-arbitrary tunability in PIMP scheme, the positions of copies should be jointly determined in the spectrum assignment algorithm. The frequencies of FWM pumps and passband of BV-WSSs should be decided based on the joint information of spectrum availability on all the multicast downstream links, so that replicas can be generated and forwarded without spectrum collision.
In conclusion, we propose a PDM-OFDM signal multicasting approach based on polarization-interleaved multi-pump FWM for all-optical elastic networks, which supports format-transparency, PDM signals, and wide spectral tunability of RO replicas. A design method for pump frequencies is given, which ensures that unwanted FWM components would not overlap usable signal replicas.
The technique has been successfully validated in experiments. An input 3-band QPSK PDM-OFDM signal is multicast in an HNLF, achieving tuning range of 1.4~1.6-THz for 4 RO-type replicas. The spacing between adjacent RO-type replicas can also be tuned from 100-GHz to 250-GHz. Finally, the scalability is demonstrated by concurrent generation of 14 replicas including 6 RO and 8 RC using only 5 pumps. The multicasting scheme offers multiple tunable RO and RC that are respectively remote from and near to the original signal in frequency domain, which provides two groups of options for replica allocation in multicasting connections. The proposed multicasting subsystem may be applied in network routing nodes to achieve more efficient link resource utilization in multicast services, e. g., IPTV, inter-datacenter service migration and data back-up [33, 37].
Funding
Program 973 (No. 2014CB340105); National Natural Science Foundation of China (NSFC) (Nos. 61505002, 61377072); CPSF (Nos. 2015M580926, 2016T90015).
References and links
1. Y. Ji, J. Zhang, Y. Zhao, H. Li, Q. Yang, C. Ge, Q. Xiong, D. Xue, J. Yu, and S. Qiu, “All optical switching networks with energy-efficient technologies from components level to network level,” IEEE J. Sel. Areas Comm. 32(8), 1600–1614 (2014). [CrossRef]
2. A. E. Willner, S. Khaleghi, M. R. Chitgarha, and O. F. Yilmaz, “All-Optical Signal Processing,” J. Lightwave Technol. 32(4), 660–680 (2014). [CrossRef]
3. B. Kozicki, H. Takara, Y. Tsukishima, T. Yoshimatsu, K. Yonenaga, and M. Jinno, “Experimental demonstration of spectrum-sliced elastic optical path network (SLICE),” Opt. Express 18(21), 22105–22118 (2010). [CrossRef] [PubMed]
4. K. Wen, X. Cai, Y. Yin, D. Geisler, R. Proietti, R. Scott, N. Fontaine, and S. J. B. Yoo, “Adaptive spectrum control and management in elastic optical networks,” IEEE J. Sel. Areas Comm. 31(1), 39–48 (2013). [CrossRef]
5. S. Chandrasekhar and X. Liu, “OFDM based superchannel transmission technology,” J. Lightwave Technol. 30(24), 3816–3823 (2012). [CrossRef]
6. G. Bosco, V. Curri, A. Carena, P. Poggiolini, and F. Forghieri, “On the performance of Nyquist-WDM terabit superchannels based on PM-BPSK, PM-QPSK, PM-8QAM or PM-16QAM subcarriers,” J. Lightwave Technol. 29(1), 53–61 (2011). [CrossRef]
7. R. Dischler, F. Buchali, and A. Klekamp, “Demonstration of bit rate variable ROADM functionality on an optical OFDM superchannel,” in Optical Fiber Communication Conference, OSA Technical Digest (Optical Society of America, 2010), paper OTuM7. [CrossRef]
8. B. Corcoran, C. Zhu, J. Schroder, L. Zhuang, B. Foo, M. Burla, W. Beeker, A. Leinse, C. Roeloffzen, and A. J. Lowery, “Multi-pass performance of a chip-enhanced WSS for Nyquist-WDM sub-band switching,” J. Lightwave Technol. 34(8), 1824–1830 (2016). [CrossRef]
9. E. Pincemin, M. Song, J. Karaki, O. Zia-Chahabi, T. Guillossou, D. Grot, G. Thouenon, C. Betoule, R. Clavier, A. Poudoulec, M. Van der Keur, Y. Jaouen, R. Le Bidan, T. Le Gall, P. Gravey, M. Morvan, B. Dumas-Feris, M. L. Moulinard, and G. Froc, “Multi-band OFDM transmission at 100 Gbps with sub-band optical switching [invited],” J. Lightwave Technol. 32(12), 2202–2219 (2014). [CrossRef]
10. S. Shimizu, G. Cincotti, and N. Wada, “Demonstration of multi-hop optical add-drop network with high frequency granular optical channel defragmentation nodes,” in Optical Fiber Communication Conference, OSA Technical Digest (Optical Society of America, 2015), paper M2I.4. [CrossRef]
11. R. Elschner, T. Richter, C. Schmidt-Langhorst, T. Kato, T. Tanimura, S. Watanabe, and C. Schubert, “Distributed aggregation of spectrally efficient single- and dual-polarization super-channels by optical frequency conversion in fiber,” J. Lightwave Technol. 34(2), 618–625 (2016). [CrossRef]
12. N. Sambo, F. Paolucci, G. Meloni, F. Fresi, L. Poti, and P. Castoldi, “Control of frequency conversion and defragmentation for super-channels [invited],” J. Opt. Commun. Netw. 7(1), A126–A134 (2015). [CrossRef]
13. R. K. Pankaj, “Wavelength requirements for multicasting in all-optical networks,” IEEE/ACM Trans. Netw. 7(3), 414–424 (1999). [CrossRef]
14. T. Rahman, G. Ellinas, and M. Ali, “Lightpath- and light-tree based groupcast routing and wavelength assignment in mesh optical networks,” J. Opt. Commun. Netw. 1(2), A44–A55 (2009). [CrossRef]
15. C.-S. Bres, N. Alic, E. Myslivets, and S. Radic, “Scalable multicasting in one-pump parametric amplifier,” J. Lightwave Technol. 27(3), 356–363 (2009). [CrossRef]
16. B. Guo, Y. Xu, P. Zhu, Y. Zhong, Y. Chen, J. Li, Z. Chen, and Y. He, “Multicasting based optical inverse multiplexing in elastic optical network,” Opt. Express 22(12), 15133–15142 (2014). [CrossRef] [PubMed]
17. D. Wang, T. Cheng, Y. Yeo, Z. Xu, Y. Wang, G. Xiao, and J. Liu, “Performance comparison of using SOA and HNLF as FWM medium in a wavelength multicasting scheme with reduced polarization sensitivity,” J. Lightwave Technol. 28(24), 3497–3505 (2010).
18. Y. Wang, C. Yu, T. Luo, L. Yan, Z. Pan, and A. E. Willner, “Tunable all-optical wavelength conversion and wavelength multicasting using orthogonally polarized fiber FWM,” J. Lightwave Technol. 23(10), 3331–3338 (2005). [CrossRef]
19. G.-W. Lu, K. S. Abedin, and T. Miyazaki, “DPSK multicast using multiple-pump FWM in Bismuths highly nonlinear fiber with high multicast efficiency,” Opt. Express 16(26), 21964–21970 (2008). [CrossRef] [PubMed]
20. M. Pu, H. Hu, H. Ji, M. Galili, L. K. Oxenløwe, P. Jeppesen, J. M. Hvam, and K. Yvind, “One-to-six WDM multicasting of DPSK signals based on dual-pump four-wave mixing in a silicon waveguide,” Opt. Express 19(24), 24448–24453 (2011). [CrossRef] [PubMed]
21. J. Wang, Q. Sun, and J. Sun, “Tunable dual-channel multicasting all-optical 40 Gbit/s logic and operation and format conversion for CSRZ signals,” Electron. Lett. 45(8), 420–421 (2009). [CrossRef]
22. G. Contestabile, M. Presi, and E. Ciaramella, “Multiple wavelength conversion for WDM multicasting by FWM in an SOA,” IEEE Photonics Technol. Lett. 16(7), 1775–1777 (2004). [CrossRef]
23. G.-W. Lu, K. S. Abedin, and T. Miyazaki, “All-optical broadband tunable wavelength multicasting using a pump modulated wide-band fiber optical parametric amplifier with high and flat gain,” in Proc. European Conference of Optical Communications (2007), pp. 1–2.
24. O. Yilmaz, S. Nuccio, X. Wu, and A. E. Willner, “Tunable N-fold multicasting and pulsewidth of 40 Gb/s channels by variable periodic slicing of a supercontinuum,” in Conference on Lasers and Electro-Optics, OSA Technical Digest (Optical Society of America, 2009), paper CTuJ5. [CrossRef]
25. A. Malacarne, G. Meloni, G. Berrettini, N. Sambo, L. Poti, and A. Bogoni, “Optical multicasting of 16QAM signals in periodically-poled lithium niobate waveguide,” J. Lightwave Technol. 31(11), 1797–1803 (2013). [CrossRef]
26. G. Meloni, V. Vercesi, M. Scaffardi, A. Bogoni, and L. Poti, “Spectral-efficient flexible optical multicasting in a periodically poled lithium niobate waveguide,” J. Lightwave Technol. 33(23), 4731–4737 (2015). [CrossRef]
27. P. Zhu, J. Li, Y. Chen, X. Chen, Z. Wu, D. Ge, Z. Chen, and Y. He, “Experimental demonstration of EON node supporting reconfigurable optical superchannel multicasting,” Opt. Express 23(16), 20495–20504 (2015). [CrossRef] [PubMed]
28. J. Lu, Z. Dong, L. Chen, and J. Yu, “Polarization insensitive wavelength conversion based on four-wave mixing for polarization multiplexing signal in high-nonlinear fiber,” Opt. Commun. 282(7), 1274–1280 (2009). [CrossRef]
29. Y. Chen, J. Li, P. Zhu, B. Guo, L. Zhu, Y. He, and Z. Chen, “Experimental demonstration of 400 Gb/s optical PDM-OFDM superchannel multicasting by multiple-pump FWM in HNLF,” Opt. Express 21(8), 9915–9922 (2013). [CrossRef] [PubMed]
30. J. Lu, L. Chen, Z. Dong, Z. Cao, and S. Wen, “Polarization insensitive wavelength conversion based on orthogonal pump four-wave mixing for polarization multiplexing signal in high nonlinear fiber,” J. Lightwave Technol. 27(24), 5767–5774 (2009). [CrossRef]
31. K. Inoue, “Four-wave mixing in an optical fiber in the zero-dispersion wavelength region,” J. Lightwave Technol. 10(11), 1553–1561 (1992). [CrossRef]
32. T. Inoue, K. Tanizawa, and S. Namiki, “Guard-band-less and polarization-insensitive tunable wavelength converter for phase-modulated signals: demonstration and signal quality analyses,” J. Lightwave Technol. 32(10), 1981–1990 (2014). [CrossRef]
33. L. Yang, L. Gong, F. Zhou, B. Cousin, M. Molnar, and Z. Zhu, “Leveraging light forest with rateless network coding to design efficient all-optical multicast schemes for elastic optical networks,” J. Lightwave Technol. 33(18), 3945–3955 (2015). [CrossRef]
34. M. Takahashi, S. Takasaka, R. Sugizaki, and T. Yagi, “Arbitrary wavelength conversion in entire CL-Band based on pump-wavelength-tunable FWM in a HNLF,” in Optical Fiber Communication Conference, OSA Technical Digest (Optical Society of America, 2010), paper OWP4. [CrossRef]
35. G.-W. Lu, T. Bo, and C. K. Chan, “Pump-phase-noise-tolerant wavelength conversion for coherent optical OFDM using coherent DFB pumping,” in Optical Fiber Communication Conference, OSA Technical Digest (Optical Society of America, 2016), paper W3D.3. [CrossRef]
36. C. Li, M. Luo, Z. He, H. Li, J. Xu, S. You, Q. Yang, and S. Yu, “Phase noise cancelled polarization-insensitive all-optical wavelength conversion of 557-Gb/s PDM-OFDM signal using coherent dual pump,” J. Lightwave Technol. 33(13), 2848–2854 (2015). [CrossRef]
37. V. Ataie, N. Alic, and S. Radic, “Optical parametric multicasting linearization based on distortion correcting tables,” IEEE Photonics Technol. Lett. 27(14), 1527–1530 (2015). [CrossRef]
