Edge-Fog-Cloud Secure Storage with Deep-Learning-Assisted Digital Twins
Zhihan Lv, Ranran Lou
Abstract
The present work studies the storage security of edge-fog-cloud computing to improve cloud storage security. The data of an intelligent manufacturing industrial machine is used in the research and pre-processed. Machine manufacturing perception data includes data storage and data transmission. In addition, combined with the perception data of machine manufacturing, digital twins technology is used to construct the digital twin in the real world to simulate the online data-driven behavior of machine manufacturing. The digital twins model is built and saved by 3Dmax. Finally, deep learning technology is introduced to defend against network intrusion. Cloud databases are vulnerable to external attacks and internal attacks of cloud data providers. Thus, the homomorphic encryption algorithm and secure multi-party computing are introduced to ensure that the database stores ciphertext data and performs data queries directly. The experimental results indicate that digital twins technology has a good effect. The two-layer cloud database model costs US$11, 476.1, the lowest among comparative models. The unencrypted Amazon Relational Database Service has the best performance. The model proposed here also achieves relatively high availability, and the hybrid cloud structure significantly improves the model's performance. The research content provides a reference for realizing data storage security in a hybrid cloud.