A Secure Architecture for Digital Twins in Resource-Constrained Industrial Systems

Digital Twins (DT) are transforming the Industrial Internet of Things (IIoT) by providing virtual replicas of physical assets, enabling enhanced predictive analytics and data-driven decision-making. However, as Industry 5.0 integrates DT with generative AI technologies via APIs and tokens, security risks increase. Outdated modeling tools, coupled with vulnerabilities in resource-constrained environments, such as limited computational power and bandwidth, further compromise DT security. These constraints impede the implementation of robust security measures like encryption, authentication, and blockchain, making DT susceptible to cyber threats and fraud. This paper addresses these challenges by proposing a secure DT conceptual architecture that leverages edge-fog-cloud computing with Federated Learning (FL), and Blockchain. The framework ensures resource-efficient security by integrating privacy-preserving computations, secure data storage, and resilient communication channels. It effectively mitigates risks beyond 51% attacks, including Sybil attacks and data manipulation, while enabling secure data exchange and realtime monitoring. Furthermore, the proposed architecture improves adaptability in complex IIoT environments, supporting seamless device integration, scalability, and energy-efficient operations to safeguard industrial systems against evolving threats.