Abstract
The FEDERALS research proposal aims to address the growing cyber threats in Critical Infrastructure (CI), such as energy, health, and government services. The digital aspect of such services has become more vulnerable due to the rapid development of Artificial Intelligence (AI) and the transition to decentralized service models. The current proposal highlights the need to develop safer and more resilient technological solutions, exploiting the potential of AI to automate threat detection and enhance security measures. The methodology that is going to be employed in this project is based on the ARCADE model, a widely accepted open architecture framework, which provides a holistic framework for the design, assessment and deployment of complex systems. Through it, all dimensions related to the operating environment of the Target System are analysed, describing its functions and its relationship with the environment, while ensuring the alignment between the business model and the technical solutions. The ARCADE model, by combining the directional viewpoint, requirements viewpoint, architectural component viewpoint, sharing viewpoint, and implementation viewpoint, provides a solid foundation for developing and implementing more secure and resilient technology solutions in response to today’s cyber threats.
The objectives of the FEDERALS project reflect a multi-dimensional approach to enhancing cyber security in critical infrastructure. Through the development of a nextgeneration platform, the project seeks to create a system that will provide robust protection against cyber-attacks, using the latest technologies in the field of encryption and anomaly detection algorithms. This platform will incorporate advanced tools and technologies, providing a comprehensive security system that can adapt to the changing threats and needs of the CIs. Federated Learning, Software-Defined Networking (SDN) and Security Traps (Honeynets) are critical elements of this effort, enabling decentralized data processing and proactive threat detection. Federated Learning allows systems to learn from data detected in different locations without requiring centralized data aggregation, while SDNs offer network flexibility and adaptability. At the same time, Honeynets act as traps for attackers, increasing the overall system security, as well as the immediate restoration of the functionalities of a CI that was attacked.
The aforementioned technologies are implemented and integrated in a platform the goal of which is also to share any data that may be generated in a potential attack, once it has been anonymised. This latter step paves the way for a sharing platform of such incidents which the community of cybersecurity lacks. In addition, this project recognises the importance of ontinuously upgrading and strengthening the robustness of the CIs. By conducting risk assessments and implementing advanced technological upgrades, the project seeks to ensure that the CIs remain one step ahead of threats, offering a high level of security and protection. In addition, the project envisages the implementation of training scenarios for personnel in critical positions through augmented and virtual reality (AR/VR) technology and the acquisition of a certificate through this technology. The proposal also provides for the standardisation and certification of security techniques and tools, with a view to promoting actions to achieve a higher level of security and confidence, as well as training and demonstration to staff, in order to inform and educate them on new technologies and secure practices.
Main Goals of FEDERALS
Development of a federated cybersecurity platform that supports Critical Infrastructures (CIs) such as healthcare, public administration, and energy networks, offering resilience against evolving cyber threats.
Continuous monitoring and response through AI-based intrusion detection and incident response systems that operate in real-time across decentralized infrastructure nodes.
Deployment of the platform in realistic CI environments, with field validation and piloting in operational settings to ensure practical applicability and stakeholder engagement.
Enhancement of national capabilities in cybersecurity, establishing Greece as a regional innovator in AI-based protection of digital public services.
Integration of Federated Learning, Software-Defined Networking (SDN), and Honeypots to enable distributed, real-time threat detection while preserving data privacy and infrastructure autonomy.
Use of Augmented and Virtual Reality (AR/VR) for the training and upskilling of cybersecurity personnel, simulating real-world attack scenarios for preparedness and mitigation.