Networking technologies are fast evolving to support the request for ubiquitous Internet access that is becoming a fundamental need for the modern and inclusive society, with a dramatic speed-up caused by the COVID-19 emergency. Such evolution needs the development of networks into disaggregated and programmable systems according to the software-defined networking (SDN) paradigm. Wavelength-division multiplexed (WDM) optical transmission and networking is expanding as physical layer technology from core and metro networks to 5G x-hauling and inter- and intra-data-center connections requiring the application of the SDN paradigm at the optical layer based on the WDM optical data transport virtualization. We present the fundamental principles of the open-source project Gaussian Noise in Python (GNPy) for the optical transport virtualization in modeling the WDM optical transmission for open and disaggregated networking. GNPy approximates transparent lightpaths as additive white and Gaussian noise channels and can be used as a vendor-agnostic digital twin for open network planning and management. The quality-of-transmission degradation of each network element is independently modeled to allow disaggregated network management. We describe the GNPy models for fiber propagation, optical amplifiers, and reconfigurable add/drop multiplexers together with modeling of coherent transceivers from the back-to-back characterization. We address the use of GNPy as a vendor-agnostic design and planning tool and as physical layer virtualization in software-defined optical networking.

GNPy model of the physical layer for open and disaggregated optical networking [Invited]

The emerging Software Defined Networking (SDN) paradigm paves the way for flexible and automatized management at each layer. The SDN-enabled optical network requires each network element’s software abstraction to enable complete control by the centralized network controller. Nowadays, silicon photonics due to its low energy consumption, low latency, and small footprint is a promising technology for implementing photonic switching topologies, enabling transparent lightpath routing in re-configurable add-drop multiplexers. To this aim, a model for the complete management of photonic switching systems’ control states is fundamental for network control. Typically, photonics-based switches are structured by exploiting the modern technology of Photonic Integrated Circuit (PIC) that enables complex elementary cell structures to be driven individually. Thus PIC switches’ control states are combinations of a large set of elementary controls, and their definition is a challenging task. In this scenario, we propose the use of several data-driven techniques based on Machine Learning (ML) to model the control states of a PIC N × N photonic switch in a completely blind manner. The proposed ML-based techniques are trained and tested in a completely topological and technological agnostic way, and we envision their application in a real-time control plane. The proposed techniques’ scalability and accuracy are validated by considering three different switching topologies: the Honey-Comb Rearrangeable Optical Switch (HCROS), Spanke-Beneš, and the Beneš network. Excellent results in terms of predicting the control states are achieved for all of the considered topologies. 

Performance evaluation of data-driven techniques for the softwarized and agnostic management of an N×N photonic switch

Optical networking is fast evolving towards the applications of the Software-defined Networking (SDN) paradigm down to the (Wavelength-division Multiplexing) WDM transport layer for cost-effective and flexible infrastructure management. Optical SDN requires each network element’s software abstraction to enable full control by the centralized network controller. Nowadays, modern network elements, especially photonic switching systems, are developed by exploiting the fast-emerging technology of Photonic Integrated Circuit (PIC) that consists of complex fabrics of elementary units that can be driven individually using a large set of elementary controls. In this work, we focus on modeling the elementary control states of the topological structures behind PIC ${N} \times {N}$ switches under a fully blind approach based on Machine Learning (ML) techniques. The ML agent’s training and testing datasets are obtained synthetically by software simulation of the photonic switch structure. The proposed technique’s scalability and accuracy are validated by considering different dimensions ${N}$ and applying it to two different switching topologies: the Honey-Comb Rearrangeable Optical Switch and the Benes network. Excellent results in terms of prediction of the control states are achieved for both of the considered topologies.

/pdf/automatic-management-of-n-n-photonic-switch-powered-by-5fd3mfx2rx.pdf

Automatic Management of N × N Photonic Switch Powered by Machine Learning in Software-Defined Optical Transport

Recent years have seen an unprecedented growth of data traffic driven by a continuous increase of connected devices and new applications. This trend will tend to saturate transparent optical networks that are the backbone of the whole telecommunication infrastructure. To improve the capacity of already deployed network infrastructures and maximize operators CAPEX returns, band-division multiplexing (BDM) has emerged as a promising solution to expand the fiber bandwidth beyond the existing C-band. Along with this, the demand for flexible and dynamically reconfigurable functionalities in each network layer is increasing. In this regard, optical networking is fast evolving towards the applications of the software-defined networking (SDN) paradigm down to the physical layer. The implementation of optical SDN requires the full abstraction and virtualization of each network element in order to enable complete control by a centralized network controller. To pursue this objective, photonics transmission components and their transmission functionalities must be abstracted to allow the definition of the control states and a real-time quality-of-transmission (QoT) evaluation of transparent lightpaths (LP). In this work, we propose an SDN based model of a photonic switching fabric that allows determining the control state and evaluating QoT degradation. Our investigations present a wideband optical switch design based on photonic integrated circuits (PICs), where QoT degradation is abstracted using a structure-agnostic approach based on machine learning (ML). The ML engine training and testing datasets are generated synthetically by software simulation of the photonic switch architecture. Results show the potential of the proposed technique to predict QoT impairments with high accuracy, and we envision its application in a real-time control plane. 

Optimized management of ultra-wideband photonics switching systems assisted by machine learning

Benes/Clos networks constitute a particularly important part of interconnection networks and have been used in numerous areas, such as multi-processor systems, data centers and on-chip networks. They have also attracted great interest in the field of optical communications due to the increasing popularity of optical switches based on these architectures. There are numerous algorithms aimed at routing these types of networks, with varying degrees of utility. Linear algorithms, such as Sun Tsu and Opferman, were historically the first attempt to standardize the routing procedure of this types of networks. They require matrix-based calculations, which are very demanding in terms of resources and in some cases involve backtracking, which impairs their efficiency. Parallel solutions, such as Lee’s algorithm, were introduced later and provide a different answer that satisfy the requirements of high-performance networks. They are, however, extremely complex and demand even more resources. In both cases, hardware implementations reflect their algorithmic characteristics. In this paper, we attempt to design an algorithm that is simple enough to be implemented on a small field programmable gate array board while simultaneously efficient enough to be used in practical scenarios. The design itself is of a generic nature; therefore, its behavior across different sizes (8 × 8, 16 × 16, 32 × 32, 64 × 64) is examined. The platform of implementation is a medium range FPGA specifically selected to represent the average hardware prototyping device. In the end, an overview of the algorithm’s imprint on the device is presented alongside other approaches, which include both hard and soft computing techniques.

/pdf/novel-benes-network-routing-algorithm-and-hardware-20173uhn.pdf

Novel Benes Network Routing Algorithm and Hardware Implementation

In order to cope with the fast increase in data traffic demand, optical networks are fast evolving towards the disaggregation and progressive implementation of the openness paradigm. Such an evolution is enabling the application of the software-defined networking below the IP layer, down to the optical transmission (SD-OTN). SD-OTN is enabled by the capability of the network controller to automatized management of photonic switching systems, and allowing their full virtualization and softwarisation. To this purpose, one of the major matter of contention is an efficient utilization of routing strategies, which can be seamlessly incorporated into the control plane. In this work, we rely on data-driven science (DDS) to develop the machine learning (ML) model which is able to predict the routing strategies of generic N x N photonic switching system without any knowledge required of the topology. The dataset used for training and testing the ML model is generated “synthetically”. In particular, the training and testing of the proposed ML module is done in a completely topological and technological agnostic way and is able to perform its application in real-time. Furthermore, the scalability and accuracy of the proposed approach is verified by considering two different switching topologies: the Honey-Comb Rearrangeable Optical Switch and the Benes network. Promising results are achieved in terms of predicting the control signals matrix for both of the considered topologies.

Machine-learning-aided abstraction of photonic integrated circuits in software-defined optical transport

We propose a general procedure for the automatic design of N x N integrated optical switches using the Benes configuration based on ring filters, suitable for multi-hop routing in meshed optical networks. The basic switch components (waveguides, couplers, optical crossings) are selected from the process development kit (PDK) of the target platform, and can be integrated with user-defined custom components. The described procedure is able to automatically select the optimal number of rings and their proper routing, minimizing the number of optical crossings and reducing the system transmission losses. The generated topology can be directly simulated in the Synopsys OptSim© environment at a system level. There, the impact of the design options on transmission performances is evaluated for the chosen signal modulation scheme and numerical simulations are automatically executed to determine, among the 2M possible optical paths across the electrically tuned M rings, the only N! configurations required to generate all the permutations of the inputs signals to the output ports minimizing the transmission BER. Finally, the generated Benes switch can be converted to a GDSII mask targeting the foundry of choice and manual adjustment to the layout can be made. The automatic procedure takes advantage of the vertical integration of the Synopsys environment, which allows to perform a comprehensive investigation of the whole photonic device, from the component design to the system level analysis, thanks to the integration of Optsim- optical communication system, Optsim schematic-driven photonic circuit, Rsoft component design tools, and OptoDesigner© mask layout.

Lorenzo Tunesi

Papers

Performance evaluation of data-driven techniques for the softwarized and agnostic management of an N×N photonic switch

Automatic Management of N × N Photonic Switch Powered by Machine Learning in Software-Defined Optical Transport

Machine-learning-aided abstraction of photonic integrated circuits in software-defined optical transport

Automatic design of NxN integrated Benes optical switch

A Neural Network-Based Automatized Management of N ×N Integrated Optical Switches