Ridge resonance has recently been discovered as a manifestation of a bound state in the continuum (BIC) on the silicon-on-insulator (SOI) platform, with the resonance achieved by the near-BIC operation controllable by the ridge width as parametric tuning. In this paper, based on the ridge resonance behavior and the coherent tunneling by adiabatic passage (CTAP) protocol, we propose and simulate a device that performs wavelength-selective mode conversion to create a resonator-like wavelength response, which can be a potential building block for an add-drop multiplexer compatible with manufacturing in silicon photonics foundry.

As a core module in elastic optical networks, reconfigurable optical add-drop multiplexers (ROADMs) are crucial for flexible resource allocation; thus, accurate and quick ROADM failure localization and identification are of paramount importance. Although machine learning (ML) holds vast potential for broad application in automated network management, differences between the training and the deployment environments, e.g., in device configurations, may lead to decreased accuracy, which poses a challenge to the quality of training data. Digital spectrum serves as effective supplementary information for soft failure localization (SFL) and identification (SFI); however, ROADM failures in wavelength division multiplexing systems can affect multiple lightpaths, complicating the strategy of employing digital spectrum. To address class imbalance for ROADM-based optical networks, this paper proposes an SFL and SFI scheme using digital spectrum based on ML. Specifically, data augmentation techniques based on the variational autoencoder (VAE) and the synthetic minority oversampling technique (SMOTE) are employed to generate synthetic samples. Subsequently, we adopt XGBoost models and leverage combinations of real and synthetic data to perform SFL and SFI. Extensive numerical analysis conducted in a four-node optical network shows that, for classes with 100× fewer data points than other classes, using SMOTE to balance the classes is more effective compared to using a VAE and yields improvements in the F1 score for localization and identification by 16.18% and 21.18%, respectively, compared to the case without data augmentation. Moreover, to simplify the model, we extract features from the digital spectrum, achieving a 1.02% improvement in the localization F1 score compared to using the digital spectrum.

This work presents a metro-access network architecture designed to meet the stringent requirements of time-sensitive applications in emerging 5G and 6G networks. The proposed architecture leverages semiconductor optical amplifier-based optical add-drop multiplexers and FPGA-based controllers to enable dynamic, low-latency operation suitable for applications with critical latency and jitter demands. We introduce a control plane protocol that allows for deterministic time-slotted resource reservation, ensuring transparent optical switching and minimizing latency. We validate the architecture through a prototype implementation in a ring network topology, demonstrating latency-bounded operation. Experimental results show that the network achieves sub-microsecond reconfiguration times and stable latency performance, with transparency crossing up to six nodes and a data recovery time of 100 ns at −17dBm, making the architecture a potential solution for federated computing and edge cloud scenarios.

Ultracompact multi-mode add-drop multiplexer based on pixelated photonic-like crystals

We present a proof-of-concept (PoC) of a metro-access optical network that uses automatically reconfigurable OpenXR Pluggables and filter-less optical add-drop multiplexers (OADMs) over a chain of access nodes connected to two metro nodes in a horseshoe topology. Each OADM consists of optical splitters and semiconductor optical amplifiers (SOAs), whose gains are individually managed by an open-source network controller. Two sets of XR Pluggables were stacked operating both in point-to-point (P2P) and point-to-multipoint (P2MP) configurations. We conducted physical-layer experiments and evaluated control plane aspects to assess the performance of the proposed network.

Open optical networks have been considered to be important for cost-effectively building and operating networks. Recently, optical-circuit-switches (OCSes) have attracted industry and academia because of their cost efficiency and higher capacity than traditional electrical packet switches and reconfigurable optical add-drop multiplexers (ROADMs). Though the open interfaces and control planes for traditional ROADMs and transponders have been defined by several standard-defining organizations, those of OCSes have not. Considering that several OCSes have already been installed in production datacenter networks and several OCS products are on the market, bringing openness and interoperability into OCS-based networks has become important. Motivated by this fact, this paper investigates a software-defined networking controller for open optical-circuit-switched networks. To this end, we identified the use cases of OCSes and derived the controller requirements for supporting them. We then proposed a multi-vendor (MV) OCS controller framework that satisfies the derived requirements; it was designed to quickly and consistently operate fiber paths upon receiving the operation requests. We validated our controller by implementing it and evaluating its performance on actual MV-OCS networks. It satisfied all the requirements, and fiber paths could be configured within 1.0 s by using our controller.

Traditionally, traffic from the access nodes is aggregated in metro-aggregation hubs over rings or horseshoes in a fixed configuration using multiple optical point-to-point (P2P) transceivers. However, this static setup limits dynamic scalability and often leads to inefficient resource usage and high costs, especially under increasing and varying traffic conditions. Coherent optical point-to-multipoint (P2MP) transceivers offer a promising solution for aggregation at this network level. These transceivers allow a single (aggregation) node to communicate with multiple (access) nodes simultaneously via digital subcarrier multiplexing (DSCM) technology. Additionally, their long-distance transmission capabilities enable the placement of the P2MP transceiver root deeper in the hierarchy. In this paper, we propose an extended access aggregation architecture that incorporates modified reconfigurable optical add–drop multiplexers (ROADMs) to support both traditional P2P and P2MP connections. This architecture allows the creation of light-trees to enable P2MP communication and hence to boost transmission flexibility and multiplexing gains. We propose a mixed-integer linear programming (MILP) model to determine the optimal placement of P2MP transceivers and the establishment of light-trees, considering physical layer impairments (PLIs) and the required quality of transmission (QoT) for the connections. To tackle the high algorithm complexity, we also introduce a best-fit decreasing heuristic to efficiently exploit the trade-off between execution time and performance. Our simulation experiments using real network topologies showcase that our proposed architecture can greatly enhance multiplexing gains while considering the operational costs tied to network expansions and upgrades.

Metro-access networks are a type of optical network connecting metro hubs with various subnetworks, covering from rural to dense urban regions. In the long term, the metro-access network is expected to address hundreds of Tb/s aggregated traffic, which makes spectral efficient multiplexing techniques a must-have. Combining wavelength division multiplexing (WDM) and digital subcarrier multiplexing (DSCM) techniques is a possible successful industrial approach to cope with this challenge. However, the ever-growing demand for bandwidth/wavelength inevitably induces increased complexity and cost of the metro-access network architecture. In this Letter, we describe the design and experimental assessment of a semiconductor optical amplifier (SOA)-based optical add-drop multiplexer (OADM) with the coherent DSCM technique, which can dynamically add and drop traffic at any node within the network and enable lossless transmission at a relatively low cost due to photonic integration. The results show that the proposed architecture can support up to five nodes at a net data rate of 291 Gb/s/λ with negligible penalty from dropping and adding operations, indicating its great potential for long-term metro-access network adoption.

This work addresses the tailoring spectral response of grating-assisted co-directional couplers (GADCs) in the context of wavelength filtering for fiber-to-the-home (FTTH) applications. Design methods for spectral response engineering by means of coupling profile apodization-type weighting techniques and also more advanced rational transfer functions fitting a predefined spectral window template are presented. Modeling results based on coupled mode theory are then applied for the design and experimental fabrication of InGaAsP/InP GADCs targeting 1.3+/1.3− µm diplexer application in FTTH access networks. The experimental results are found to be in good agreement with the modeling predictions. The design tools presented are quite general and can be easily adapted to other technology platforms, such as silicon photonics for the use of GADCs as add-drop wavelength division multiplexers. The field of parity–time symmetry is another avenue where these types of gain–loss-assisted GADCs as active components are of interest for switching applications, and the design methods presented here may find utility.

Terahertz is one of the most promising technologies for high-speed communication and large-scale data transmission. As a classical optical component, ring resonators are extensively utilized in the design of band-pass and frequency-selective devices across various wavebands, owing to their unique characteristics, including optical comb generation, compactness, and low manufacturing cost. While substantial progress has been made in the study of ring resonators, their application in terahertz surface wave systems remains less than fully optimized. This paper presents several spoof surface plasmon polariton-based devices, which were realized using ring resonators at terahertz frequencies. The influence of both the radius of the ring resonator and the width of the waveguide coupling gap on the coupling coefficient are investigated. The band-stop filters based on the cascaded ring resonator exhibit a 0.005 THz broader frequency bandwidth compared to the single-ring resonator filter and achieve a minimum stopband attenuation of 28 dB. The add–drop multiplexers based on the asymmetric ring resonator enable selective surface wave outputs at different ports by rotating the ring resonator. The devices designed in this study offer valuable insights for the development of on-chip terahertz components.

The estimation of spectral spacing (guard band) among optical channels in gridless WDM systems would be decisive for making swift decisions in reconfigurable optical add-drop multiplexers (ROADMs) to avoid linear interchannel interference (ICI) effects during the channel aggregation process in transit nodes. In this work, we propose a method based on the construction of heat scatter images from constellation diagrams along with convolutional neuronal networks (CNN) to identify when optical channels are spectrally overlapped as well as the value in GHz of the channel separation in a specific optical channel without adjacent channels information. We validate our method in a gridless 16-QAM Nyquist-WDM system with different channel spacing and optical signal-to-noise ratio (OSNR). Experimental results demonstrated that overlap detection can be achieved with 3×16G B d accuracy. Additionally, the estimation of spectral spacing achieved an error (RMSE) of less than 0.6GHz. The penalty in accuracy and RMSE is only ∼98% and , respectively, when there is no knowledge of the OSNR value. Thus, this method has the potential to be integrated into monitoring tools designed for future dynamic gridless optical transmission systems.

We present in this paper a detailed brainstorming on the future option of merging the metro and the passive optical network (PON) access network segments, enabled by the introduction of end-to-end coherent transmission. We begin by reporting the experimental results presented by our group at OFC2024 (for which this paper is an invited extension). Starting from these preliminary but very promising results, we elaborate on two different possible schematics for metro+PON convergence using edge reconfigurable optical add-drop multiplexers (ROADMs) at the boundary of the two segments, and then we study their physical layer scalability by a mix of experimental characterization and numerical modeling. We show that coherent transceivers enable excellent performance in this scenario, allowing at least 200G per wavelength and even 400G in most cases when traversing all-optically two ROADMs before being routed towards a high splitting ratio PON in the access part of the network. We study several realistic conditions analyzing different bit rates, modulation formats, and network architectures, showing the physical layer conditions that would enable the PON optical distribution network loss to be in the range from 29 to 35 dB, as required by current international standards. The scalability analysis is first based on link budget and optical signal-to-noise ratio (OSNR) fundamental limitations, and it is then extended considering other physical layer issues, such as tight optical filtering in the ROADMs.

This paper discusses an optical continuum as a key technique to help telecom operators reduce the total cost of ownership (TCO) of their optical networks. Enabling photonic technologies are reviewed, with a focus on recent advances in reconfigurable optical add drop multiplexers (ROADMs) based on silicon photonics.

Optical switch is an essential component in integrated photonic circuits. A mode-conversion nanobeam cavity (MCNC) coupled with two waveguides has been employed to realize ultra-compact and low-loss 2 × 2 thermo-optic switches in silicon-on-insulator. This system can exhibit either Fano or Lorentzian lineshape in transmission spectra, dependent on the coupling structure. It has a low dropping loss, and two outputs are in the same direction, owing to the unidirectional coupling between the resonant mode and bus waveguides. Here, we have demonstrated a high-performance 2 × 2 Fano switch with a bandwidth of 5.2 nm and a footprint of only 35.5 × 1 µm2. The insertion loss (IL) and crosstalk (CT) are 0.7 dB and −54.1 dB in the bar state, respectively, while the IL and CT are 0.9 dB and −17.4 dB in the cross-state, respectively. In addition, 2 × 2 optical switches with a Lorentzian transmission lineshape have also been realized and then applied to construct a four-channel reconfigurable optical add-drop multiplexer (ROADM). Through thermal tuning, the ROADM has achieved a channel spacing of 200 GHz or 400 GHz, with an inter-channel CT below −12.3 dB or −17.2 dB, respectively. To our knowledge, we have reported the first demonstrations of 2 × 2 Fano switch and ROADM based on MCNCs. The proposed 2 × 2 switches will find potential applications in advanced photonic integrated circuits.

Scalable methods for optical transmission performance prediction using machine learning (ML) are studied in metro reconfigurable optical add-drop multiplexer (ROADM) networks. A cascaded learning framework is introduced to encompass the use of cascaded component models for end-to-end (E2E) optical path prediction augmented with different combinations of E2E performance data and models. Additional E2E optical path data and models are used to reduce the prediction error accumulation in the cascade. Off-line training (pre-trained prior to deployment) and transfer learning are used for component-level erbium-doped fiber amplifier (EDFA) gain models to ensure scalability. Considering channel power prediction, we show that the data collection process of the pre-trained EDFA model can be reduced to only 5% of the original training set using transfer learning. We evaluate the proposed method under three different topologies with field deployed fibers and achieve a mean absolute error of 0.16 dB with a single (one-shot) E2E measurement on the deployed 6-span system with 12 EDFAs.

Addressing the capacity, low cost, and low power challenges of 6G distribution networks, this paper proposes and demonstrates a multi-band optical metro-access network architecture employing semiconductor optical amplifier (SOA)-based wavelength division multiplexing (WDM) switches as low-cost and low-power multi-band optical add-drop multiplexers (MB-OADMs) to extend the capacity beyond C-band limits. We implemented and evaluated the performance of an SOA-based MB-OADM prototype in the C- and O-bands including the network reconfigurability, the node scalability, and the capability to support high-capacity transmission. Experimental results show that the MB-OADM-based network maintains high optical signal-to-noise ratio (OSNR) values up to 35.38 dB in the C-band and 33.56 dB in the O-band over 100 km across five nodes with a 20 km linkspan in between (a total of 100 km) without additional optical amplifiers at 25 Gbps. This work also assesses the MB-OADM-based network scalability in terms of nodes and data rate. Results indicated that the architecture supports cascading through nine C-band nodes over 45 km with a 3.6 dB power penalty at 25 Gbps for a bit error rate (BER) of 10−6 and through four O-band nodes with a 3 dB penalty at 25 Gbps for a BER of 10−6, maintaining nearly uniform power levels across channels at a BER under the FEC threshold. It successfully demonstrates PAM-4 at 50 Gbps and 100 Gbps data rate transmission operation crossing four nodes over a 4 km distance, with 2 dB and 2.4 dB power penalty at a BER of 10−3 in the C-band and 1.85 dB and 2.2 dB power penalty at a BER of 10−3 in the O-band, respectively.

A variable optical attenuator (VOA) is a crucial component for optical communication, especially for a variable multiplexer (VMUX) and reconfigurable optical add-drop multiplexer (ROADM). With the capacity increasing dramatically, a large-port-count and low-power-consumption VOA array is urgent for an on-chip system. In this paper, we experimentally demonstrate a 16-channel VOA array based on a polymer/silica hybrid waveguide. The proposed array is able to work over C and L bands. The VOA array shows an average attenuation larger than 14.38 dB with a low power consumption of 15.53 mW. The low power consumption makes it possible to integrate silica-based passive devices with a large port count on-chip.

Integrated switches play a crucial role in the development of reconfigurable optical add-drop multiplexers (ROADMs) that have greater flexibility and compactness, ultimately leading to robust single-chip solutions. Despite decades of research on switches with various structures and platforms, achieving a balance between dense integration, low insertion loss (IL), and polarization-dependent loss (PDL) remains a significant challenge. In this paper, we propose and demonstrate a 32 × 4 optical switch using high-index doped silica glass (HDSG) for ROADM applications. This switch is designed to route any of the 32 inputs to the express ports or drop any channels from 32 inputs to the target 4 drop ports or add any of the 4 ports to any of the 32 express channels. The switch comprises 188 Mach-Zehnder Interferometer (MZI) type switch elements, 88 optical vias for the 44 optical bridges, and 618 waveguide-waveguide crossings with three-dimensional (3D) structures. At 1550 nm, the fiber-to-fiber loss for each express channel is below 2 dB, and across the C and L bands, below 3 dB. For each input channel to all 4 drop/add channels at 1550 nm, the loss is less than 3.5 dB and less than 5 dB across the C and L bands. The PDLs for all express and input channels to the 4 drop/add channels are below 0.3 dB over the C band, and the crosstalk is under −50 dB for both the C and L bands.

The development of optical networks in the telecommunications sector is becoming much closer by considering the help of an Optical Add Drop Multiplexer (OADM) based on a novel technology called Wavelength Division Multiplexing (DWDM). The objective of the current study to examine high transmission bit rates for next-generation optical communication networks using the technology of OADM Based on DWDM. Artificial neural networks (ANNs) were developed via MATLAB software to predict three main parameters in this filed such as transmitted signal power (PT), transmitted signal bandwidth (B.Wsig), and transmission bit rate capacity (Bsh) at different fiber cable lengths, such as L=200, 250, and 300 km. The ANNs results showed that, standard error (SE) for predicting PT as a function of the number of transmitted channels (Nch) was 0.115 mW, 0.095 mW and 0.077 mW, for 200, 250 and 300 km, respectively. Additionally, the SE for predicting B.Wsig was 0.067 GHz, 0.051 GHz and 0.040 GHz for 200, 250 and 300 km, respectively. Lastly, the SE for predicting Bsh was 1.665, 1.311 Gbit/sec and 1.076Gbit/sec for 200, 250 and 300 km, respectively. The SE for predicting PT as a function of the Signal Wavelength (λ) was 0.116, 0.096 and 0.079 mW for 200, 250 and 300 km, respectively. Additionally, the SE for predicting B.Wsig was 0.067, 0.052 and 0.052 GHz for 200, 250 and 300 km, respectively. Lastly, the SE for predicting Bsh was 1.688, 1.417 and 1.110 Gbit/sec for 200, 250 and 300 km, respectively. The low SE in ANNs demonstrated the efficiency, motivating further advancements in optimizing network performance for high-bit-rate transmission.

Metro-access networks exploiting wavelength division multiplexing (WDM) to cope with the ever-growing bandwidth demands are sensitive to cost and need to be fast-configurable to meet the requirements of many new network services. Optical add-drop multiplexers (OADMs) are a key component in enabling fast dynamic wavelength allocation and optimization. In this Letter, we propose and demonstrate, to our knowledge, a novel architecture for high-performance metro-access networks that utilizes semiconductor optical amplifier (SOA)-based OADM nodes, digital subcarrier multiplexing (DSCM), low-cost direct detection receivers, and power loading techniques, which makes the designed metro-access network cost-effective, fast reconfigurable, and flexible for bandwidth allocation on demand. Through a proof-of-concept experiment, we have successfully demonstrated a prototype horseshoe optical network consisting of up to four SOA-based OADM nodes at 40 Gb/s per wavelength channel by leveraging the proposed scheme. The flexible bandwidth allocation and dynamic add and drop operations have also been achieved in an emulated WDM optical network. All results indicate the great scalability and flexibility of the proposed architecture.