Open Access
Add to BibSonomyAbstract
We experimentally demonstrate the feasibility of a multi-degree colorless, directionless, and contentionless (C/D/C-less) ROADM node composed of high port count wavelength-selective switches and transponder aggregators using silica-based planar lightwave circuit technology. The experimental results show that the introduction of a C/D/C-less function to a multi-degree ROADM node induces no significant penalty in a 127-Gbit/s PDM-QPSK signal transmission.
©2011 Optical Society of America
1. Introduction
Internet traffic has been increasing greatly over the last few years because of the spreading use of various new services such as video on demand and cloud computing. To deal with this growth in traffic, optical transport networks are evolving to realize more flexibility as regards wavelength routing and wavelength assignment. As optical transport networks move toward highly dynamic mesh-based topologies, flexible optical node architecture has become vital for minimizing resource requirements. Multi-degree reconfigurable optical add/drop multiplexing (ROADM) with a colorless, directionless and contentionless (C/D/C-less) function provides numerous benefits, because any add/drop port can support all colors and connect to any degree [1]. For example, the C/D/C-less ROADM reduces the complexity of network planning, enables spare transponders to be shared among all the wavelength channels in the node and facilitates optical path protection/restoration for link or node failures. Figure 1 shows the 8-degree C/D/C-less ROADM node architecture. Several node configurations have already been proposed for implementing a C/D/C-less function. As regards colorless switches for connecting input and output express paths (hereafter referred to as wavelength cross-connects; WXC), they have a common structure composed of a wavelength-selective switch (WSS) and a power splitter. On the other hand, there are a number of different approaches for achieving a directionless and contentionless function, such as using a WSS, a combination of splitters, switches and tunable filters, and matrix switches [2, 3].
At the European Conference on Optical Communication held in 2011, we proposed a C/D/C-less ROADM node configuration consisting of a high port count WSS and transponder aggregators (TPA) using silica-based planar lightwave circuit (PLC) technology, and reported experimental results showing its feasibility for a 127-Gbit/s polarization-division-multiplexing quadrature phase-shift keying (PDM-QPSK) signal transmission system [4]. In this paper, we describe the design concept of a C/D/C-less ROADM node configuration in more detail. In addition, we explain why multi-channel detection using a digital coherent receiver with balanced detection imposes a considerable penalty. We were unable to explain this satisfactorily in [4] owing to the space limitation. The rest of this paper is organized as follows. In Section 2, we describe our proposed node configuration and explain its superiority to the widely reported configuration. Then, Section 3 examines the penalty resulting from the multi-channel detection needed in our proposed node configuration, where the TPA does not have a wavelength-channel selection function. Section 4 presents experimental results showing that the introduction of a C/D/C-less function to a multi-degree ROADM node imposes no significant penalty on a 100G digital coherent detection system. Section 5 concludes this paper.
2. C/D/C-less ROADM node configuration
As shown in Fig. 1, the C/D/C-less ROADM node is composed of a WXC, a TPA and a transmitter/receiver (Tx/Rx). Figure 2 shows the optical signal path connection between the WXC, the TPA and the Tx/Rx. On the drop side (blue line in Fig. 2), the incoming wavelength-division-multiplexing (WDM) signals from the degree-1 fiber are delivered to other add-side WXCs or dropped to the drop-side TPA by way of the drop-side WXC-1. The dropped signals from all the drop-side WXCs are collected and allocated to the desired transponders by the TPA. On the add side (red line in Fig. 2), the output signal from the transponder is routed to the desired add-side WXC by the add-side TPA. The add-side WXC-1 collects added signals and signals from other drop-side WXCs, and then launches them into the degree-1 output fiber.
Here, let us discuss the number of optical components. A WXC composed of 1 × p switch can be connected to p-M + 1 TPAs. M denotes the number of degrees, which is assumed to be eight in this work. Figure 3 shows examples of the WXC configuration. Figure 3(a) corresponds to a widely reported configuration using commercially available components, that is, a 1 × 8 splitter and a 1 × 9 WSS. In this instance, one and two TPAs are connected to the drop and add-side WXCs, respectively. To reduce the node loss, the 1 × 8 splitter with an intrinsic splitting loss of 9 dB could be replaced with a 1 × 9 WSS as shown in Fig. 3(b). Figure 3(c) shows our proposed WXC configuration composed of two 1 × 43 WSSs [5], where up to thirty-six TPAs can be connected. When the TPA has M input and q output ports (M × q switch), the node can admit a total of (p-M + 1) × q transponders. Figure 4 shows the add/drop ratio as a function of the port count of the WSS in the WXC for an 8-degree 80-wavelength WDM system as an example. Note that a node with a 100% add-drop ratio can accommodate 640 ( = 8 degree × 80 channel) transponders. This result clearly shows that the adoption of a higher port count WSS in the WXC is an appropriate way to achieve a higher add/drop ratio. Of course, the concatenation of a 1 × 9 WSS makes it possible to increase the add/drop ratio. However, this approach has major drawbacks such as increases in the insertion loss and the required number of components. In this work, we experimentally verified the feasibility of our proposed C/D/C-less ROADM node by using 1 × 43 WSSs and 8 × 8 TPAs as an example. The results are described in Section 4.
Fig. 3 WXC configuration for using (a) 1 × 8 splitter and 1 × 9 WSS (conventional), (b) 1 × 9 WSS, (c) 1 × 43 WSS (our proposal).
Next, we explain how to realize a C/D/C-less ROADM node with a 100% add/drop ratio (full add/drop node). Figure 5 shows the path connection between the WXC, the TPA and the Tx/Rx for the full add/drop node. Compared with the path connection shown in Fig. 2, to provide a full add/drop node, additional splitters and combiners are required for the drop and add sides, respectively. Figure 6 shows examples of the WXC and the splitter/combiner combination. It is clear that the higher port count WSS in the WXC contributes to the reduction of the splitter/combiner scale, which leads to a reduced insertion loss. The total loss including that of the WXC, the splitter/combiner and the TPA for the add/drop path is described at the end of this section.
Fig. 6 WXC and splitter/combiner configuration for using (a) 1 × 8 splitter and 1 × 9 WSS (conventional), (b) 1 × 9 WSS, (c) 1 × 43 WSS (our proposal).
Figure 7 shows the drop-side TPA configuration for q = 8. In Fig. 7(a), the WSS delivers the WDM signals dropped from one drop-side WXC to each desired 1 × 8 switch with respect to each wavelength, and switches connect the path between the WSS and the transponder. On the other hand, in Fig. 7(b), the dropped WDM signals are broadcasted to all the WSSs by a 1 × 8 splitter. The WSS selects the desired wavelength-channel signal from the desired degree for the transponder. This “splitter and WSS” configuration (Fig. 7(b)) has certain drawbacks and advantages compared with the “WSS and switch” configuration (Fig. 7(a)). The “WSS and switch” configuration has serious disadvantages in terms of size, cost and power consumption due to its huge number of switching components. In addition, a much larger number of power monitor points must be installed in the TPA to specify the location of faulty switches. In contrast, the “splitter and WSS” configuration has a larger insertion loss due to the splitting loss, and this drawback becomes more severe as the output port number, q, increases. We think that broadcasting by a 1 × q splitter is a more realistic solution as long as no additional amplifiers are needed to compensate for the splitting loss as regards the design of the loss level diagram for the drop path. Figure 7(c) shows our proposed TPA configuration consisting of 1 × 8 splitters and switches. The significant benefit is that all the components in one TPA can be integrated in one chip using photonic integration technologies [6]. In particular, we believe that a silica-based PLC switch could reduce redundancy against component failures because of its field-proven reliability. In Fig. 7(c), since we select a desired wavelength channel by tuning the frequency of the local oscillator (LO) in the coherent receiver instead of using a tunable filter, we can eliminate the wavelength selection function in the TPA, and this contributes to a substantial reduction in the required number of optical components. However, the resulting receiver overload and power crosstalk penalty must be considered. The evaluation results for the multi-channel detection penalty are reported in Section 3.
Fig. 7 TPA configuration for using (a) WSS and switch, (b) splitter and WSS, (c) splitter and switch (one-chip integrated PLC).
Finally, we discuss the add/drop-path losses for a combination consisting of the above-mentioned WXC and TPA configurations. Table 1 shows the estimated losses for the combination comprising the WXC in Fig. 3 and the TPA in Fig. 7. We calculated the add/drop-path losses based on the losses of the WSS and the switch being 5 and 2 dB, respectively. Note that the WSS in the add-side TPA is not indispensable and can be replaced with a switch without a wavelength selection function because the transponder outputs only one wavelength-channel signal. However, for simplicity, we assumed that the add-side TPA has the same configuration as the drop side TPA. Table 2 shows the losses when we combine the WXC in Fig. 6 (full add/drop node) and the TPA in Fig. 7. For the full add/drop node, additional amplifiers will be required to compensate for the loss of the splitter/combiner, because the minimum signal power at the input of a digital coherent receiver is specified at −18 dBm in an implementation agreement released by the Optical Internetworking Forum (OIF) [7]. In particular, a higher power or more amplifiers will be needed for the conventional WXC with the 1 × 9 WSS and the large scale splitter/combiner configuration (Fig. 6(a) and 6(b)) compared with our proposed WXC using the 1 × 43 WSS (Fig. 6(c)). The feasibility of implementing an additional amplifier in the add/drop path is an important study item.
Table 1. Insertion Loss for Add/Drop Path
Table 2. Insertion Loss for Full-Add/Drop Node
The advantages of our proposed C/D/C-less ROADM node configuration are summarized as follows. As regards the WXC configuration shown in Fig. 3, the higher port count WSS in the WXC can provide the higher add/drop ratio without an additional component and loss increase as shown in Fig. 4 and Table 1. For the full-add/drop node composed of the WXC and the splitter/combiner shown in Fig. 6, the higher port count WSS reduces the splitter/combiner scale and add/drop path losses, which lead to the reduction of the required output power or number of the additional amplifiers. As regards the TPA, we believe that the integration of the splitter and switch array using silica-based PLC technology substantially reduces the size and cost, thus, our integration approach becomes more realistic solution to provide multi-degree C/D/C-less ROADM node.
3. Multi-channel detection penalty
In our proposed TPA as shown in Fig. 7(c), a desired wavelength channel is selected by tuning the LO frequency in the coherent receiver instead of using an optical filter. Therefore, up to q wavelength-channel signals are launched into the receiver simultaneously. Thus, the resulting receiver overload and power crosstalk penalty must be considered. Although the optical signal-to-noise ratio (OSNR) penalties for 40-Gbit/s PDM-QPSK signal detection using a single-ended coherent receiver have been investigated [8], neither the penalties for a 127-Gbit/s system based on balanced detection nor an approach for suppressing them has yet been reported. Thus, we experimentally examined the multi-channel detection penalty for a 127-Gbit/s PDM-QPSK signal.
First, we explain why multi-channel detection using a digital coherent receiver with balanced detection induces a considerable penalty. Reference [8] points out that the crosstalk signal power is expected to be eliminated from the output of a coherent receiver with balanced detection. We believe that the multi-channel detection penalty in balanced detection is caused by an imperfection in the coherent receiver, namely, an insufficient common mode rejection ratio (CMRR). Figure 8(a) shows the schematic configuration of a 90-degree optical hybrid in a coherent receiver.
Fig. 8 Schematic configuration for (a) 90-degree optical hybrid in coherent receiver, (b) CMRR for signal port, (c) CMRR for LO port.
The CMRR for the signal input port is written as
where aS+ and aS- are the splitting ratio of a 3-dB coupler (aS+2 + aS-2 = 1), and R+ and R- are the responsivity of the photodiode (PD) as shown in Fig. 8(b). In an ideal receiver, aS+ 2 = aS-2 = 0.5 and R+ = R-, that is, CMRRS = 0. Similarly, the CMRR for the LO input port (Fig. 8(c)) is expressed asThe CMRR in a coherent receiver is described in greater detail in [9]. The output currents △I of the balanced PD for the in-phase (△II) and quadrature (△IQ) components are written as follows:In Eq. (3), PS and PL are the powers of the input signal and LO light, respectively. θt is the phase of the signal in reference to the LO light phase, θt = θS(t) - θL(t). △IDC is an unwanted output caused by the insufficient CMRR and written asHere, we consider the output from the balanced PD for multi-channel detection. The WDM signals launched into the signal input port are expressed aswhere N is the number of wavelength channels, and PXTi and θXTi are the power and phase of the i-th channel signal, respectively. Note that the amplified spontaneous emission (ASE) noise component is disregarded because its power is negligible compared with the signal or LO light powers. In this case, the output currents are written as follows:This equation means that the output currents from the balanced PD are disturbed by the crosstalk channel power owing to the insufficient CMRR for the signal port.Next, we report our experimental result for the multi-channel detection penalty. Multi-channel (up to 8) 127-Gbit/s PDM-QPSK signals in the C-band were launched into a coherent receiver without optical filtering, and only one channel signal was demodulated by tuning the LO frequency. The received signal power was −10 dBm/ch and the OSNR with a resolution of 0.1 nm was adjusted to 23 dB. Figure 9 shows the Q-factor penalty as a function of the received channel number. The LO power was changed from 7 to 13 dBm. This result indicates that the multi-channel detection penalty can be substantially suppressed by increasing the LO power. We think this is because the crosstalk power effect given by Eq. (6) exhibits a relative decrease when we increase the LO power, which is given by PL in Eq. (3).
4. Experimental demonstration of multi-degree C/D/C-less ROADM
We verified the C/D/C-less function of our proposed ROADM node experimentally. First, we ensured that the WXC and TPA worked properly as a component. Figure 10(a) and 10(b) show the experimental setup and results for directionless function verification. In Fig. 10(a), the path from the Tx side to one WXC (1 × 43 WSS) was switched from Tx #1 to #8. On the other hand, in Fig. 10(b), the path from one Tx to the WXC side was switched from WXC #1 to #8. To be precise, the TPA outputs are connected to separate WXCs in a normal situation. However, in our experiment, the TPA outputs are connected to separate ports of the same 1 × 43 WSS for simplicity. This simplification has no effect on the verification of the TPA function. Figure 10(c) corresponds to contentionless function verification. All the Txs launched the same wavelength signals into the TPA simultaneously, and the TPA routed one signal (Tx #4) to the WXC from #1 to #8, while the other signals were delivered to the remaining WXC. In all the experiments, we changed the signal wavelength from 1534.25 to 1565.09 nm to ensure the colorless function over the entire C-band. We measured Q-factors with a received OSNR of 17 dB adjusted by the ASE source. The measured Q-factors are plotted at the bottom of Fig. 10. We compared these Q-factors with that obtained in an experiment with a back-to-back configuration without the TPA and WSS, and confirmed that there are no penalties for any of the paths in the TPA.
Fig. 10 Experimental results for colorless WXC and TPA directionless (input side) (b) directionless (output side) (c) contentionless function.
Next, we report the results of a transmission experiment representing a single-direction 8-degree ROADM. Figure 11 shows the concept of our transmission experiment for directionless function verification. The WDM signals from the Tx were delivered to a transmission path composed of a 20-km single-mode fiber and an erbium-doped fiber amplifier (EDFA) to compensate for the span loss. The add-side TPA and WXC switched the sending path from path I to Path IV in series (Fig. 11(a)–11(d)). In response to the path change, the drop-side TPA and WXC switched the receiving path to deliver the received WDM signals to the Rx. Figure 12 (a) –12(d) show the path setting in the node for the four transmission paths shown in Fig. 11(a)–11(d). Four 127-Gbit/s PDM-QPSK signals with a 100-GHz spacing were generated at the Tx. The add-side TPA and WSS sent these signals to Paths I ~IV. Then, four signals were launched into the coherent receiver by way of the drop-side TPA and WSS, and each signal was demodulated by tuning the frequency of the LO with an output power of 13 dBm. The received signal power was adjusted to −10 dBm/ch with the EDFA in each transmission path. Figure 13 shows the measured Q-factors. The dashed line corresponds to the results for a single channel transmission, where the TPAs were removed.
Figures 14 and 15 show the concept and path setting used for contentionless function verification. Four signals at the same wavelength were launched into the add-side TPA, and traveled through different paths. The four signals reached the drop-side TPA at the same time. The TPA guided these signals to different Rxs. Figure 16 shows the measured Q-factors. The results shown in Fig. 13 and Fig. 16 indicate that the introduction of a C/D/C-less function with the multi-channel detection technique causes no significant penalties for a 127-Gbit/s PDM-QPSK signal transmission system.
5. Conclusion
We described a C/D/C-less ROADM node configuration composed of 1 × 43 WSSs and PLC-based TPAs, and explained its advantage over the commonly suggested configuration. A higher port count WSS in the WXC can provide a higher add/drop ratio without any additional components or loss increase. Alternatively, it reduces the splitter/combiner scale and add/drop path losses for the full add/drop node. As regards the TPA, the one-chip integration of the splitter and switch array substantially reduces the size and cost. In addition, we experimentally examined the Q-factor penalty resulting from the multi-channel detection needed in our proposed node configuration. Finally, we reported the results of a transmission experiment showing that the introduction of a C/D/C-less function to a multi-degree ROADM node caused no significant penalty for a 127-Gbit/s PDM-QPSK signal transmission system.
References and links
1. S. Gringeri, B. Basch, V. Shukla, R. Egorov, and T. J. Xia, “Flexible architectures for optical transport nodes and networks,” IEEE Commun. Mag. 48(7), 40–50 (2010). [CrossRef]
2. R. Jensen, A. Lord, and N. Parsons, “Colourless, directionless, contentionless ROADM architecture using low-loss optical matrix switches,” ECOC 2010, Mo.2.D.2 (2010).
3. S. Nakamura, S. Takahashi, M. Sakauchi, T. Hino, M. B. Yu, and G. Q. Lo, “Wavelength selective switching with one-chip silicon photonic circuit including 8×8 matrix switch,” OFC/NFOEC 2011 OTuM2 (2011).
4. Y. Sakamaki, T. Kawai, T. Komukai, M. Fukutoku, T. Kataoka, T. Watanabe, and Y. Ishii, “Experimental demonstration of colourless, directionless, contentionless ROADM using 1×43 WSS and PLC-based transponder aggregator for 127-Gbit/s DP-QPSK system,” ECOC 2011, Th.12.A.3 (2011).
5. Y. Ishii, K. Hadama, J. Yamaguchi, Y. Kawajiri, E. Hashimoto, T. Matsuura, and F. Shimokawa, “MEMS-based 1×43 wavelength-selective switch with flat passband,” ECOC 2009, PD 1.9 (2009).
6. T. Watanabe, K. Suzuki, T. Goh, K. Hattori, A. Mori, T. Takahashi, T. Sakamoto, K. Morita, S. Sohma, and S. Kamei, “Compact PLC-based transponder aggregator for colorless and directionless ROADM,” OFC/NFOEC 2011, OTuD3 (2011).
7. Optical Internetworking Forum, “Implementation agreement for integrated dual polarization intradyne coherent receivers,” (2010). http://www.oiforum.com/public/documents/OIF_DPC_RX-01.0.pdf.
8. L. E. Nelson, S. L. Woodward, S. Foo, M. Moyer, D. J. S. Beckett, M. O’Sullivan, and P. D. Magill, “Detection of a single 40 Gb/s polarization-multiplexed QPSK channel with a real-time intradyne receiver in the presence of multiple coincident WDM channels,” J. Lightwave Technol. 28(20), 2933–2943 (2010). [CrossRef]
9. Y. Painchaud, M. Poulin, M. Morin, and M. Tetu, “Performance of balanced detection in a coherent receiver,” Opt. Express 17(5), 3659–3672 (2009). [CrossRef] [PubMed]
