Achieving Privacy-Preserving Sensitive Attributes for Large Universe Based on Private Set Intersection

Abstract Increasingly more user data are transferred to the cloud for storage and analysis. Due to the dishonesty of the cloud, user data are at risk of leakage. CP-ABE is considered one of the most promising technologies for protecting outsourced data. However, in most existing schemes, the attacker may learn user privacy information from the policy plaintext. Furthermore, schemes supporting policy hiding either only achieve partial policy hiding or cannot hide policies in large universes. In this paper, we use hide both the values and names of attributes in policies by using private set intersections (PSIs). CP-ABE supports a complete hiding strategy. At the same time, it can calculate the authorization relationship and mapping between the user’s password and key. We use polynomial-based PSI and a recursive algorithm to calculate the former. With the help of this algorithm and tag vectors, the mapping is determined in the case of communication restrictions. By outsourcing exponents, we achieve an efficient hiding policy and effectively reduces users’ computing overhead. We also prove its security. Finally, the performance evaluation and simulation results reveal that our model achieves better performance compared with other schemes while preserving sensitive attributes.