Intent-Based Networking (IBN) has emerged as a promising paradigm in network management, enabling the translation of high-level user intents into automated network configurations. To fully realize the potential of IBN, Artificial Intelligence (Al) plays an important role in enhancing the efficiency, adaptability, and scalability of its core building blocks. This paper explores the role and impact of AI across the main functionalities of IBN, including intent recognition and translation, intent validation and optimization, real-time assurance, and closed-loop feedback systems. We detail how techniques such as Natural Language Processing (NLP), Machine Learning (ML), and Reinforcement Learning (RL) are leveraged to automate and refine each phase of the IBN process. Additionally, we examine the benefits of AI integration with IBN, such as improved network performance, reduced operational complexity, and predictive capabilities, while addressing challenges like interpretability, computational overhead, and data privacy. This study highlights the critical contributions of AI to advancing IBN frameworks and provides a roadmap for future innovations in this field.
Programmable photonic integrated circuits have the potential to increase the speed at which photonic applications are developed. However, in order to use these circuits effectively, there is a lack of efficient algorithms that compute an appropriate configuration for a given use case. We discuss a place -and -route algorithm that quickly calculates a solution to this problem. A specialized multi -stage simulated annealing process is introduced that gives a fast and effective way to configure the port -to -port response of programmable photonic integrated circuits.
The paper explores advanced sensing technologies for improving the security and resilience of optical networks. It introduces systems like Distributed Acoustic Sensing (DAS) and polarization analysis, enabling accurate anomaly detection and real-time monitoring, which assist in identifying and preventing security breaches. The paper also examines how machine learning analyzes complex patterns to optimize detection processes. Practical experiments and simulations demonstrate the effectiveness of these technologies in safeguarding critical infrastructures, including fiber manipulation detection and environmental diagnostics.
Fibre sensing is undergoing a resurgence of interest due to the potential of integrating it into the telecommunications infrastructure and leveraging this infrastructure to achieve scale and widespread applicability. This is a trend that is occurring in both wireless and wired/fibre systems. Through this approach, the backbone fibre networks can be turned into a massive sensor for detecting earthquakes, tsunamis and a host of geographic disturbances [1], [2]. On a metro scale, networks have also been shown to detect the flow of road traffic and transport systems [3] using both the coherent communication signals themselves and a variety of sensing probes. These methods are largely compatible with high data rate dense wavelength division multiplexed (DWDM) networks and a large volume of research is carried out to improve the sensitivity and performance of the sensing technologies [4]. While promising, these early experiments remain proof of concept tests, in which the sensing technologies are a bolted-on novelty. Full exploitation of the benefits of the integrated sensing and communication (IASC), such as improved performance, increased security of the infrastructure and the creation of new services, requires sensing that is not just an add-on service, but a key part of the operation and performance of the communication system. This means a control and management of a sensing system that is compatible with optical network control and management, allowing for sensing signals to be deployed and routed across the network similar to the data channels, and the intelligence gained from the sensing system informing the operation of the communications network.
The performed numerical analysis demonstrates an application of a surface plasmon phenomenon for efficient modulation of light intensity in the reflection mode. A supply of modulating high frequency electrical signal to an electro-optical thin layer located upon a thin metallic layer supporting surface plasmons is assumed. The narrow plasmonic and waveguide dips in the angular reflection spectrum are sensitive to the change of the permittivity of the cover electro-optical nano-layer. Numerical analysis is performed for the relevant metal-dielectric multi-nano-layer structure at the oblique incidence of transverse magnetic (TM) wave. For electromagnetic analysis the frequency-domain method of single expression (MSE) is used. The performed analysis permitted to obtain all necessary parameters relevant to the thicknesses of metallic and electro-optic material for efficient optical wave intensity modulation.
We evaluate the convergence of 4 G and 6 G in an optical access network for a fibre link of 25 Km. The signals were transmitted alone LTE at 1.5 GHz and OFDM-IM in baseband. An optimum power of -21 dBm was found for LTE. Then the distance between LTE and OFDM-IM was optimized, achieving a minimum distance of 129 MHz, beside the optimum power of -20 dBm for LTE was achieved when were transmitted together and a 4 dBm improvement in the power receiver for LTE was found.
In this paper, we compare the equalization performance of three different passive reservoir computing architectures using micro-ring resonators as reservoirs. The architectures consist of four resonators, arranged in a parallel 4x1, sequential 1x4 and deep 2x2 configuration. The reservoirs are simulated using the two-dimensional finite-difference time-domain method and are fed with simulation data from a comprehensive simulation setup with realistic component characteristics. We show, that the parallel architecture outperforms both the sequential and the deep micro-ring resonator configurations. Moreover, the parallel structure with a linear readout outperforms state-of-the-art digital signal processing in terms of bit error ratio by up to an order of magnitude. This passive reservoir computing architecture could pave the way to more energy-efficient and high symbol rate intra-DCN and mobile fronthaul systems.
Parallel transmission based on OFDM is undoubtedly an excellent solution for frequency-selective and bandlimited channels. For general linear channels with additive Gaussian noise, the channel capacity can be achieved in combination with water pouring and ideal coding. In conjunction with optical transmission based on intensity modulation and direct detection (IM/DD), however, non-negative signals are required, so that the matter is no longer so simple. Here, the actual OFDM signal is usually provided with a DC-offset that corresponds to the mean optical power to be used. Since signals always tend to be Gaussian distributed in parallel transmission, the information bearing component is preferably clipped before biasing, or an adaptive bias is used. However, what about alternative solutions? What are the benefits of Kramers-Kronig optical OFDM or of “single side-band OFDM” under the IM/DD constraint? What if we clip only a single modulated subcarrier (that has lower peak values) symmetrically, apply a bias and equalize the received signal in the frequency domain? In this context we consider carrierless amplitude-phase (CAP) modulation, where the bandpass pulses are generated directly by means of FIR-filtering. Of course, “unfiltered” pulse-amplitude modulation could also be an alternative.
In this contribution, we present and discuss two results selected from these problems, with the concentration given to the different special modeling techniques, we have recently developed. First, we have developed a specific numerical technique - magnetooptic 3D MO aperiodic Rigorous coupled-wave analysis (aRCWA), in order to cope successfully with the complexity of wave effects in photonic nonreciprocal (one-way) structures, based on MO effects where unidirectional propagation can appear in the presence of an external magnetic field. We will first report here on a specific extension of the RCWA/aRCWA tool, a fully anisotropic method (including general form of anisotropy of both permittivity and permeability) in three dimensional (3D) case, due to its major application also called magnetooptic RCWA/aRCWA (MORCWA/MOaRCWA). We have then focused on the investigation of the spectral and transmission properties of structures based on InSb/dielectric waveguides at THz frequencies, in particular, on a 3D InSb hybrid Si plasmonic channel guiding structure. Both dispersion diagrams of surface magnetoplasmons, as well as relative spectral transmittances of the forward and backward propagating waves have been provided for this structure, proving the one-way behavior. The method can be easily applied for simulating and designing novel MO (and even fully bi-anisotropic) applications, in more complex novel configurations resembling realistic designs.
Reservoir computing is a recurrent computing architecture particularly suited for implementation on photonic hardware, especially in its time-delay (time-multiplexed) implementation. Here, we review our effort in timedelay reservoir computing based on a single nonlinear node consisting of a silicon microring resonator. We focus on relating the physical effects taking place in the resonator to the computing performance, as well as proposing extensions of the standard scheme to allow for increased computing efficiency, e.g. through task-multiplexing, or reduced footprint, e.g. through input-multiplexing.
This research considers the evolution of optical networks and the rising difficulties linked to managing optical signal switching within the network, especially when dealing with multi-band over spatial division multiplexing (MBoSDM) systems. We put forward three unique node architectures. Each node’s complexity is scrutinized and compared with a standard node architecture, offering valuable perspectives on potential obstacles in future optical networks. The work further investigates the advantages of employing fibre-core switching, a method that capitalizes on the reality that multiple spatial paths will link the same nodes, and traffic volume may allow to avoid individual channel demultiplexing/multiplexing at every node of the network. The results of this study not only enhance our comprehension of node architectures and fibre-core switching but also lay the groundwork for more effective and resilient optical communication networks.
Sommerfeld-Zenneck wave is a long-wavelength asymptotic form of a surface wave at a metal-dielectric boundary. Away from the SPP resonance, the Sommerfeld-Zenneck mode is weakly confined and does not feature decreased phase or group velocities. Still coupling a Sommerfeld-Zenneck wave in the vicinity of absorbing materials contributes to enhanced absorption. This paper focuses on the interaction between Sommerfeld-Zenneck and guided modes in a type-II superlattice high operating temperature long-wavelength infrared photodetector. Owing to enhanced absorption in the active layer, it is possible to retain the original level of absorptance and to reduce the absorber thickness down to a few hundred nanometers. Coupling surface and guided modes at similar resonance frequencies is possible thanks to a two-dimensional sub-wavelength hole array in a gold film integrated with the detector. The coupling mechanism may be approximately explained based on a semi-analytical model dependent on the dispersion relations of the main detector components partly obtained from ellipsometric measurements. Yet 3D vectorial optical modelling going down to deeply sub-wavelength dimensions is needed in a refined approach. Excitation of hybrid modes provides a mechanism for enhancing the detector’s quantum efficiency and tailoring the detector’s response at specific resonance frequencies.
We discuss how the design of heterogeneously integrated laser diodes can be optimized to guarantee high wall-plug efficiency for those lasers as well as make them very insensitive to external (parasitic) reflections. We also provide some experimental results obtained for heterogeneously integrated DFB lasers on silicon. We also introduce an opto-electronic feedback loop that allows to modify the external feedback to give optimum linewidth and side mode rejection.
As some physical parameter measurements (e.g. vibration) require a high-speed FBG demodulation system, a high-speed scanning module as the core part of this system is investigated. In this study, the design and implementation of a high-speed scanning module for Fiber Bragg Grating (FBG) sensing based on a Distributed Bragg Reflector (DBR) tunable laser and a Field Programmable Gate Array (FPGA) is presented. The innovatively designed and constructed high-speed FBG sensor system comprises a DBR laser, a laser drive circuit with self-programmed control algorithms, and a signal processing module. The overall performance of this high-speed FBG sensor system is evaluated and verified experimentally in terms of the validation of scanning rate, accuracy and stability. This high-speed FBG sensor system has opened a wide range of unique applications for the measurement of a variety of key physical parameters which require a high-speed FBG demodulation technique.
This paper illustrates the design of triple layer transmitarray, in planar shape configuration, having phase modulation capability for beam steering application at f = 27 GHz. A transmitarray unit cell, comprising three layers composed of identical circular ring slots is designed. Each layer is separated by a distance t = 0.254 mm. The aforesaid unit cell is employed as a basis element for a 13x13 planar transmitarray with 2-bit monofocal phase compensation capability. The transmitarray is fed by a standard horn antenna having an approximate gain of 10 dBi with a focal distance of F = 44.7 mm. Simulation results show that planar transmitarray allows a phi(BS) = +/- 19-degree beam steering capability with a A(SL) = 2 dB scan loss and A(SLL) = -10.6 dB side lobe level. About 10 dB of gain increment with respect to the standard double-ridged horn antenna is obtained for the planar transmitarray, respectively.
In this work, we propose the use of sensors based on fiber Bragg gratings for identifying the relative position of walls in large facilities and corridors, for example in mines. In smart large facilities such as mines, a backbone network is usually included to automate tasks and maintain safety. By connection of the fiber optic sensor detection network with wireless and radio networks, location information regarding the walls can be sent to the main computer managing the large facility. In this way, safety in large industrial facilities can be improved.
The 1980s saw networks begin a transition from copper-cable moderated growth to an acceleration of creativity powered by optical fibre and mobile technology. The widespread digital connectivity of people and machines led to mobile working, social media, eCommerce, Cloud Computing, and over-the-top (OTT) media services. Throughout this era, network providers failed to anticipate, and satisfy, the manifest growth potential despite demonstrably solid Moore's Law (related) predictions. They also failed to recognise mounting energy costs and the coming crisis posed by the expansion of cloud computing, networked Artificial Intelligence (AI), Internet-of-Things (IoT), 5/6G, and (possibly) quantum computing! In this session we review the status quo and propose some radical solutions to the likely energy demands of a growing IoT.
In this paper, we show the idea of a non-volatile multilevel memory, realized with an optical resonator. The cross-section of the resonator is realized with Si3N4 buried in SiO2, with segmented layers of Sb2Se3 periodically deposited above the Silicon Nitride (separated by a thin layer of Silicon Dioxide). Thanks to the programmability of each segment of Sb2S3 by means of thermal pulses, a digitalized average effective refractive index along the resonator is obtained. In this way it is possible to obtain a multilevel switch with amplitude reading around a wavelength of 1550 nm.
The provisioning of deterministic services over E2E multi-technological infrastructures requires the proper configuration of the resources to match the required Key Performance Indicators (KPI), such as latency and jitter. From an architectural point of view, such configuration is carried out by an SDN-based TSN Controller that has the responsibility to control multiple technologies. In this paper, the main challenges of the TSN controller in such scenario are deeply discussed and its role and implementation in the TIMING architecture is also discussed.
