We present the first elastic, space division multiplexing, and multi-granular network based on two 7-core MCF links and four programmable optical nodes able to switch traffic utilising the space, frequency and time dimensions with over 6000-fold bandwidth granularity. Results show good end-to-end performance on all channels with power penalties between 0.75 dB and 3.7 dB.
© 2013 OSA
Space Division Multiplexing (SDM) transmission technology has been proposed as a viable solution for increasing the transmission capacity of optical fibres by exploiting additional channels (e.g. cores, modes) in the spatial domain . With the continuous and ever growing development of improved multi-core fibres, coupling devices and multi-core amplifiers, SDM promises to deliver higher capacity, reduced footprint and eventually lower costs than conventional single mode fibre systems. However, in order to further exploit this technology in optical networking, it is necessary to introduce new optical node and network architectures that support the increased number of spatial channels per fibre, and yet are able to switch traffic at diverse granularities, e.g. fibre/core, waveband, wavelength and sub-wavelength switching. They should also provide transparent interoperability between heavily-loaded network areas with high-capacity Multi-Core Fibre (MCF) links and areas with conventional Single Mode Fibre (SMF) links. Similarly, the requirement for future optical transport networks able to carry mixed bitrates, e.g. 10 Gb/s, 100 Gb/s, 400 Gb/s, 1 Tb/s and beyond, has triggered a great deal of interest in elastic optical networks. In such networks spectrum allocation is performed in a flexible manner, depending on the requirements of individual channels . Several solutions have been proposed to improve scalability or flexibility of optical nodes. Multi-granular optical cross-connects (MG-OXC) aim to improve scalability by grouping several wavelengths into bands and switching them with a single cross-connection . However, MG-OXCs do not support elastic spectrum allocation and provide limited support for evolving requirements and new functionalities such as sub-wavelength switching, wavelength conversion, regeneration and other forms of signal processing. Meanwhile, an optical node based on bandwidth-variable wavelength selective switches (BV-WSS) in a broadcast and select arrangement has been proposed to support elastic spectrum allocation . However, the very nature of this architecture will restrict scalability and upgradability to space division multiplexing, and will limit support for evolving requirements and new functionalities. Recently proposed large-scale optical cross-connect architectures [5,6] provide high scalability but limited support for elastic spectrum allocation and new functionalities. Therefore, we propose and demonstrate here an SDM multi-dimensional switching network that supports high bandwidth flexibility, mixed traffic, conventional single core and new multi-core fibres by leveraging on programmable Architecture on Demand (AoD) optical nodes incorporating varying degrees of networking functionality.
In this paper, we present results from the first optical multi-dimensional networking demonstration based on two 7-core MCFs and four programmable AoD all-optical nodes able to switch traffic in space, frequency and time, with a range of bandwidth granularity of over 6000 fold . We successfully demonstrate end-to-end transport of 5.7-Tb/s traffic, using a combination of fixed/flexgrid, elastic band and sub-wavelength switching. Channels include 4x555 Gb/s, 60x42.7 Gb/s and 38x10 Gb/s wavelength channels, plus 12x42.7 Gb/s and 6x10 Gb/s time-multiplexed sub-wavelength.
2. Multi-granular switching using space, frequency and time
Optical networks using SDM technology will be required to switch large volumes of traffic combined with the flexibility to switch bands or individual wavelength and sub-wavelength channels. In this paper we demonstrate that by switching in space, frequency and time, as shown in Fig. 1(a), it is possible to provide a large range of all-optical granularities in networks with SDM transmission technology. When the space domain is used, entire fibre /cores are switched without demultiplexing the signals they carry. Therefore, there is no need for (de)multiplexing devices, which improves node scalability. However, fibre/core switching is only possible if all the signals in a given input go to the same output. If this is not the case, fibres/cores may require demultiplexing into individual bands or channels and aggregated at the output, e.g. using spectrum selective switches (SSSs). Moreover, spectrum slots can be further divided into time slices, to support time-multiplexed sub-wavelength traffic.
3. Architecture-on-demand nodes
In a dynamic network, where switching requirements may change over time, a flexible network infrastructure is required in order to cope effectively with such changes. Hence, we propose using AoD nodes, which are able to adapt and evolve with traffic requirements . As illustrated in Fig. 1(c), an AoD node consists of an optical backplane, e.g. based on a large port-count 3D-MEMS , connected to the node’s inputs and outputs and several plug-in modules that provide the required signal processing functions, e.g. SSSs , fast switches , EDFAs, wavelength converters, regenerators, etc. Based on traffic requirements, an AoD synthesis algorithm calculates a synthetic node design that provides the required functionality with the available modules. Such synthesis algorithms are presented elsewhere . Furthermore, if traffic requirements change an alternative synthetic node design may be calculated and implemented in order to fulfill the new requirements. The time required to implement a new synthetic node depends on the speed of the optical backplane and the number of cross-connections to configure. However, once a synthetic node has been setup the optical backplane speed has no influence on the supported bit rates as cross-connections remain unchanged until a new synthetic node, with different functionalities, is required.
An illustrative example of a synthetic node design that supports multiple switching granularities is shown in Fig. 1(b). It provides fibre/core switching granularity from input C to output G. Elastic wavelength and waveband granularity from input A to output E are implemented with an SSS. Also, sub-wavelength switching granularity is implemented with a PLZT fast switch (e.g. 10 ns switching time), which is set to ON during time-slots occupied by sub-wavelength channels that need to be passed through to the output, and OFF elsewhere. Although at the PLZT input there is a copy of all signals from input A, and they all undergo the same alternating ON/OFF process, the undesired signals are filtered out by the SSS. Thus, only the spectral components of the required sub-wavelength channels is passed through to output E. Synthetic node designs, such as the one shown in Fig. 1(b), are implemented by creating appropriate cross-connections in the optical backplane, so that fibre inputs/outputs and modules are interconnected to construct the required synthetic node. For instance, the cross-connections depicted in Fig. 1(c) would implement the synthetic node design presented in Fig. 1(b).
One major benefit of AoD is the flexibility to support arbitrary switching granularities on any port. Various switching granularities are implemented by selectively introducing modules within the synthetic node design that provide the required switching function, e.g. (DE)MUX or Wavelength Selective Switch (WSS) devices for fixed-grid switching, SSSs for flexgrid/gridless switching, fast switches for time-based sub-wavelength switching, etc. Also, due to their efficient multi-granular support, AoD nodes can provide substantial scalability gains. High scalability is achieved by switching traffic at the coarsest granularity, i.e. fibre/core switching, so that single backplane cross-connections are able to switch large volumes of traffic. At the same time, the system is able to provide switching at finer granularities (e.g. fixed/flexgrid, subwavelength switching) plus additional functionality when necessary, with extra modules inserted only when and where required. Therefore, AoD can yield overall hardware reductions .
This is illustrated in Fig. 2(a), where all channels from input cores 2 and 7 require switching to output cores 6 and 7 respectively. In order to fulfill signals’ switching requirements, an AoD node utilizes only two cross-connections. In contrast, a conventional static ROADM broadcasts the channels to all possible outputs only to select them again at the required output, as shown in Fig. 2(b). Hence, in this case the conventional ROADM approach introduces higher loss and is less hardware efficient than the AoD approach, as the modules employed to implement the broadcast and select functions (optical amplifiers, power splitters and SSSs) are underutilized. Also, as illustrated in Fig. 2(a), the AoD node is able to provide additional functionality when required, e.g. wavelength conversion. It is possible to achieve hardware reductions by dimensioning the number of modules that provide additional functionality according to their specific demand, so that the blocking probability is lower than a given threshold. For instance, if the demand for a specific functionality increases a higher number of modules that provide such functionality will be required (connected to the backplane) in order to maintain the same blocking probability. This method of dimensioning additional functionalities (services) in an AoD node is enabled by the fact that modules can be connected freely to any input/output ports, thereby making it possible to share modules. In contrast, in a static node design sharing modules is not always possible as they are fixed to specific inputs/outputs. Therefore, static nodes will often require a higher number of modules than AoD nodes to achieve the same blocking probability for a specific additional functionality.
4. MCF links
As shown in Fig. 3(a), a 2-km trench-assisted 7-core fibre (MCF-1), and a 3-km single-step index homogeneous 7-core fibre (MCF-2) are used for connecting nodes 1-3. Inset A of Fig. 3(a) shows the MCF facets. Both MCFs were fabricated by Mitsubishi Cable Industries, and connected to optical nodes using SDM MUX/DEMUX devices based on free-space optics . There are two main contributions to the total core loss of the MCF links, one due to the multi-core fibre itself and the other due to the free-space coupling optics. The total loss of the MCF and SDM MUX/DEMUX was measured for each core in both MCF links, results are presented in Fig. 4(a). The average loss across all cores was 2 dB for MCF-1 and 2.4 dB for MCF-2. The maximum core loss fluctuation observed over an 11-hour period was 0.4 dB and 0.2 dB for MCF-1 and MCF-2 respectively, as shown in Fig. 4(b). The average measured crosstalk (including MUX/DEMUX contributions) was −56.5 dB and −53.8 dB for MCF-1 and MCF-2 respectively, as depicted in Fig. 4(c).
5. Experimental SDM and multi-granular network setup and results
The elastic multi-dimensional switching network is comprised of four programmable optical nodes of different sizes and capabilities, as shown in Fig. 3(a). Nodes 1-4 are based on the AoD concept  and consist of an optical backplane that interconnects MCF/SMF fibre inputs, functional modules and MCF/SMF fibre outputs, as shown in Fig. 1(c). The optical backplanes of Nodes 1, 2 and 4 are implemented as independent partitions of a 160x160 3D-MEMS optical switch with a 20-ms switching time. Node 3 is implemented with a 16x16 beam steering switch . Average losses of both switches per cross-connection are 2 dB and 0.59 dB, respectively. Node-4 is linked to Nodes 1, 2 and 3 by different lengths of SMF, shown in Fig. 3(a).
Signals A-J, which are input to nodes 1-4, consist of mixtures of 555 Gb/s Discrete Multi-tone (DMT) , 42.7 Gb/s, and 10 Gb/s OOK NRZ signals (continuous or sub-wavelength channels) with different destinations as described in Table 1. Various lengths of fibres, including an 80-km installed fibre between the towns of Colchester and Ipswich, are inserted for signal transmission or de-correlation, as shown in Fig. 3(a). Fibre/core switching is realized in Node-1, where SMF inputs B-E, G, and K are 〉exibly connected to MCF-1 core 1-6. Fixed-grid switching is also demonstrated in Node-1, where 9x10G signals from H and 4x10G signals from Node 4 (I) are combined in a 200-GHz Arrayed Waveguide Grating (AWG) and transmitted over core 7 of MCF-1 to Node-2. Switching across multiple dimensions with the largest granularity range is demonstrated in Node-2. The core/core switching has the coarsest granularity, and the signals from cores 1-4 of MCF-1 are directly switched to cores 1-4 of MCF-2 because they have common destinations (Node 3). On the other hand, signals from core 7 and inputs A, F and J (Figs. 3(b)i-iv) consist of channels with different destinations. Such spatial fragmentation was effectively solved with the AoD node functions. Figure 3(b)v-vii show the spectra of Node-2 outputs, where 700 GHz elastic band switching and 50G-600G 〉exgrid switching are demonstrated by means of a Liquid Crystal on Silicon (LCOS)-based Spectrum Selective Switch (SSS) . Figure 3(b)viii represents an elastic sub-wavelength switching of 10G/42.7G channels using a 10-ns PLZT switch , where time-slot size is variable within 1-18µs on a 64-µs frame. Since the minimum data rate of sub-wavelength switched 10G channel and maximum data rate of a space-switched core amount to 156 Mb/s and 939.3 Gb/s respectively, these results demonstrate 6000-fold elastic bandwidth granularity.
Figure 5(a) shows the end-to-end OSNR of the channels after ampli□cation and dispersion compensation. BER results for the 42.7G signals that traverse the longest path (Tx-2 F λ12-λ18 and λ22-λ27) showed a maximum penalty at BER = 10−9 of 2.4 dB, Fig. 5(b). End-to-end BER curves for Tx-1 A λ0 (555G) are presented in Fig. 5(c). All sub-carriers showed BER under 2x10−3 for average received powers higher than −32 dBm. Figure 5(d) shows typical BER curves where penalties of Tx-1 A λ0, Tx-1 B λ3, Tx-2 F λ18 and Tx-1 A sub-λ2 were 1.75 dB, 0.75 dB, 1.7 dB and 3.7 dB (due to high PLZT loss). The measured inter-core crosstalk penalty for 555G and 42.7G over MCF1-core1 and MCF2-core1 was negligibly small, as shown in Fig. 6.
We have presented the ﬁrst SDM multi-granular switching network based on two 7-core MCFs and four programmable AoD all-optical nodes able to switch trafﬁc with a range of over 6000-fold elastic bandwidth granularities utilizing the space, frequency and time dimensions. We demonstrated sub-wavelength, ﬁxed/ﬂexgrid, elastic band, space switching and spatial defragmentation of 5.7-Tb/s trafﬁc with good end-to-end performance.
This work is supported by the EPSRC grant EP/I01196X: The Photonics Hyperhighway. The authors are grateful to Mitsubishi Cable Industries from Japan for providing the MCFs and would like to thank Polatis for the loan of the low-loss space switch and Yenista Optics for the bandwidth-variable optical ﬁlter.
References and links
1. J. Sakaguchi, B. J. Puttnam, W. Klaus, Y. Awaji, N. Wada, A. Kanno, T. Kawanishi, K. Imamura, H. Inaba, K. Mukasa, R. Sugizaki, T. Kobayashi, and M. Watanabe, “19-core fiber transmission of 19x100x172-Gb/s SDM-WDM-PDM-QPSK signals at 305Tb/s,” in Optical Fiber Communication Conference. Optical Society of America, 2012, p. PDP5C.1.
2. O. Gerstel, M. Jinno, A. Lord, and S. J. B. Yoo, “Elastic optical networking: a new dawn for the optical layer?” IEEE Commun. Mag. 50(2), s12–s20 (2012). [CrossRef]
3. X. J. Cao, V. Anand, Y. Xiong, and C. Qiao, “A study of waveband switching with multilayer multigranular optical cross-connects,” IEEE J. Sel. Areas Comm. 21(7), 1081–1095 (2003). [CrossRef]
4. M. Jinno, H. Takara, B. Kozicki, Y. Tsukishima, Y. Sone, and S. Matsuoka, “Spectrum-eﬃcient and scalable elastic optical path network: architecture, benefits, and enabling technologies,” IEEE Commun. Mag. 47(11), 66–73 (2009). [CrossRef]
5. T. Ban, H. Hasegawa, K. Sato, T. Watanabe, and H. Takahashi, “A novel large-scale OXC architecture that employs wavelength path switching and fiber selection,” in European Conference and Exhibition on Optical Communication, Optical Society of America, 2012, p. We.3.D.1.
6. Y. Iwai, H. Hasegawa, and K. Sato, “Large-Scale photonic node architecture that utilizes interconnected small scale optical cross-connect sub-systems,” in European Conference and Exhibition on Optical Communication, Optical Society of America, 2012, p. We.3.D.3.
7. N. Amaya, M. Irfan, G. Zervas, R. Nejabati, D. Simeonidou, J. Sakaguchi, W. Klaus, B. J. Puttnam, T. Miyazawa, Y. Awaji, N. Wada, and I. Henning, “First fully-elastic multi-granular network with space/frequency/time switching using multi-core fibres and programmable optical nodes,” in European Conference and Exhibition on Optical Communication. Optical Society of America, 2012, p. Th.3.D.3.
8. N. Amaya, G. S. Zervas, B. R. Rofoee, M. Irfan, Y. Qin, and D. Simeonidou, “Field trial of a 1.5 Tb/s adaptive and gridless OXC supporting elastic 1000-fold bandwidth granularity,” in Optical Communication (ECOC), 2011 37th European Conference and Exhibition on, Sept. 2011.
9. M. C. Wu, O. Solgaard, and J. E. Ford, “Optical MEMS for lightwave communication,” J. Lightwave Technol. 24(12), 4433–4454 (2006). [CrossRef]
10. G. Baxter, S. Frisken, D. Abakoumov, H. Zhou, I. Clarke, A. Bartos, and S. Poole, “Highly programmable wavelength selective switch based on liquid crystal on silicon switching elements,” in Optical Fiber Communication Conference, 2006 and the 2006 National Fiber Optic Engineers Conference. OFC 2006, March 2006.
11. K. Nashimoto, D. Kudzuma, and H. Han, “High-speed switching and filtering using PLZT waveguide devices,” in OptoeElectronics and Communications Conference (OECC), 2010 15th, July 2010, 540 –542.
12. M. Garrich, N. Amaya, G. S. Zervas, P. Giaccone, and D. Simeonidou, “Architecture on demand: Synthesis and scalability,” in Optical Network Design and Modelling (ONDM), 2012 16th Conference on, Apr. 2012.
13. W. Klaus, J. Sakaguchi, B. J. Puttnam, Y. Awaji, N. Wada, T. Kobayashi, and M. Watanabe, “Free-space coupling optics for multi-core fibers,” in Photonics Society Summer Topical Meeting Series, 2012 IEEE, July 2012, 230 –231.
14. T. A. Truex, A. A. Bent, and N. W. Hagood, “Beam-steering optical switch fabric utilizing piezoelectric actuation technology,” http:/www.polatis.com/products/technology.asp.
15. N. Amaya, M. Irfan, G. Zervas, K. Banias, M. Garrich, I. Henning, D. Simeonidou, Y. R. Zhou, A. Lord, K. Smith, V. J. F. Rancano, S. Liu, P. Petropoulos, and D. J. Richardson, “Gridless optical networking field trial: Flexible spectrum switching, defragmentation and transport of 10G/40G/100G/555G over 620-km field fiber,” in 37th European Conference and Exhibition on Optical Communication (ECOC), 2011, Sept. 2011.