Research on Cloud Data Security Computing Framework Based on Fusion of Homomorphic Encryption and Differential Privacy

With the wide application of cloud computing in network security, the privacy protection of sensitive data is becoming increasingly serious. This paper proposes a cloud data security computing framework that combines homomorphic encryption and differential privacy. It supports ciphertext computing based on the CKKS scheme, and introduces ε-differential privacy mechanism at the output end to achieve “invisible in calculation and unrecognizable after calculation” Double protection. Based on UNSW-NB15 and CERT v6.2 datasets, the experiment carries out intrusion detection and behavior aggregation tasks respectively. Under the privacy budget ε=1.0, the F1-score of intrusion detection task reaches 92.3%, and the re-recognition rate decreases to 6.7%; The behavioral aggregation error is controlled within 1.92%, which is better than baseline methods such as HE-only and DP-only. The results show that the framework can significantly improve the level of privacy protection while ensuring data availability. It is suitable for various scenarios such as intrusion detection and anomaly modeling, and has strong practicability and promotion value.