Network Digital Twin (NDT) is a new technology that builds on the concept of Digital Twins (DT) to create a virtual representation of the physical objects of a telecommunications network. NDT bridges physical and virtual spaces to enable coordination and synchronization of physical parts while eliminating the need to directly interact with them. There is broad consensus that Artificial Intelligence (AI) and Machine Learning (ML) are among the key enablers to this technology. In this work, we present B5GEMINI, which is an NDT for 5G and beyond networks that makes an extensive use of AI and ML. First, we present the infrastructural and architectural components that support B5GEMINI. Next, we explore four paradigmatic applications where AI/ML can leverage B5GEMINI for building new AI-powered applications. In addition, we identify the main components of the AI ecosystem of B5GEMINI, outlining emerging research trends and identifying the open challenges that must be solved along the way. Finally, we present two relevant use cases in the application of NDTs with an extensive use of ML. The first use case lays in the cybersecurity domain and proposes the use of B5GEMINI to facilitate the design of ML-based attack detectors and the second addresses the design of energy efficient ML components and introduces the modular development of NDTs adopting the Digital Map concept as a novelty.

The rapid evolution of network infrastructure through the softwarization of network elements has led to an exponential increase in the attack surface, thereby increasing the complexity of threat protection. In light of this pressing concern, European Telecommunications Standards Institute (ETSI) TeraFlowSDN (TFS), an open-source microservice-based cloud-native Software-Defined Networking (SDN) controller, integrates robust Machine-Learning components to safeguard its network and infrastructure against potential malicious actors. This work presents a comprehensive study of the integration of these Machine-Learning components in a distributed scenario to provide secure end-to-end protection against cyber threats occurring at the packet level of the telecom operator’s Virtual Private Network (VPN) services configured with that feature. To illustrate the effectiveness of this integration, a real-world emerging attack vector (the cryptomining malware attack) is used as a demonstration. Furthermore, to address the pressing challenge of energy consumption in the telecom industry, we harness the full potential of state-of-the-art Green Artificial Intelligence techniques to optimize the size and complexity of Machine-Learning models in order to reduce their energy usage while maintaining their ability to accurately detect potential cyber threats. Additionally, to enhance the integrity and security of TeraFlowSDN’s cybersecurity components, Machine-Learning models are safeguarded from sophisticated adversarial attacks that attempt to deceive them by subtly perturbing input data. To accomplish this goal, Machine-Learning models are retrained with high-quality adversarial examples generated using a Generative Adversarial Network.

Cybercrime has become more pervasive and sophisticated over the years. Cyber ranges have emerged as a solution to keep pace with the rapid evolution of cybersecurity threats and attacks. Cyber ranges have evolved to virtual environments that allow various IT and network infrastructures to be simulated to conduct cybersecurity exercises in a secure, flexible, and scalable manner. With these training environments, organizations or individuals can increase their preparedness and proficiency in cybersecurity-related tasks while helping to maintain a high level of situational awareness. SPIDER is an innovative cyber range as a Service (CRaaS) platform for 5G networks that offer infrastructure emulation, training, and decision support for cybersecurity-related tasks. In this paper, we present the integration in SPIDER of defensive exercises based on the utilization of machine learning models as key components of attack detectors. Two recently appeared network attacks, cryptomining using botnets of compromised devices and vulnerability exploit of the DoH protocol (DNS over HTTP), are used as the support use cases for the proposed exercises in order to exemplify the way in which other attacks and the corresponding ML-based detectors can be integrated into SPIDER defensive exercises. The two attacks were emulated, respectively, to appear in the control and data planes of a 5G network. The exercises use realistic 5G network traffic generated in a new environment based on a fully virtualized 5G network. We provide an in-depth explanation of the integration and deployment of these exercises and a complete walkthrough of them and their results. The machine learning models that act as attack detectors are deployed using container technology and standard interfaces in a new component called Smart Traffic Analyzer (STA). We propose a solution to integrate STAs in a standardized way in SPIDER for the use of trainees in exercises. Finally, this work proposes the application of Generative Adversarial Networks (GANs) to obtain on-demand synthetic flow-based network traffic that can be seamlessly integrated into SPIDER exercises to be used instead of real traffic and attacks.

Recently, a novel approach in the field of Industry 4.0 factory operations was proposed for a new generation of automated guided vehicles (AGVs) that are connected to a virtualized programmable logic controller (PLC) via a 5G multi-access edge-computing (MEC) platform to enable remote control. However, this approach faces a critical challenge as the 5G network may encounter communication disruptions that can lead to AGV deviations and, with this, potential safety risks and workplace issues. To mitigate this problem, several works have proposed the use of fixed-horizon forecasting techniques based on deep-learning models that can anticipate AGV trajectory deviations and take corrective maneuvers accordingly. However, these methods have limited prediction flexibility for the AGV operator and are not robust against network instability. To address this limitation, this study proposes a novel approach based on multi-horizon forecasting techniques to predict the deviation of remotely controlled AGVs. As its primary contribution, the work presents two new versions of the state-of-the-art transformer architecture that are well-suited to the multi-horizon prediction problem. We conduct a comprehensive comparison between the proposed models and traditional deep-learning models, such as the long short-term memory (LSTM) neural network, to evaluate the performance and capabilities of the proposed models in relation to traditional deep-learning architectures. The results indicate that (i) the transformer-based models outperform LSTM in both multi-horizon and fixed-horizon scenarios, (ii) the prediction accuracy at a specific time-step of the best multi-horizon forecasting model is very close to that obtained by the best fixed-horizon forecasting model at the same step, (iii) models that use a time-sequence structure in their inputs tend to perform better in multi-horizon scenarios compared to their fixed horizon counterparts and other multi-horizon models that do not consider a time topology in their inputs, and (iv) our experiments showed that the proposed models can perform inference within the required time constraints for real-time decision making.