Efficient Non-Subset Membership Encryption for Privacy-Preserving Project Collaborative Evaluation

Zhenhua Chen
College of Computer Science and Technology, Xi’an University of Science and Technology, Xi’an, 710054, China
Zhian Guan
College of Computer Science and Technology, Xi’an University of Science and Technology, Xi’an, 710054, China

Abstract:

Project Collaborative Evaluation includes evaluating a project and making a decision on whether to go on to the next stage of this project or give up it. In order to accomplish this task, therefore, it is necessary to set up an evaluation team to meet a certain requirement that at least one of members is not from the project company so as to ensure the objectivity of the evaluation. However, the project data during evaluation, often as commercial information of companies, are quite confidential and usually unable to be revealed to anybody except the evaluation team whose members are subject to the restriction as required. Accordingly, how to conduct such an evaluation without compromising the privacy of project data has become a crucial issue. Motivated by this issue, in this paper we first introduce a new cryptographic concept called non-subset membership encryption (NSME). In this concept, the project data specified as M encrypted with a set of members from the project company denoted as S, can be decrypted under a decryption key generated with a set of evaluators denoted as W, if W is not a subset of S which means all the evaluators don’t come from a same project company. For the sake of practicality, we then provide a high-efficiency NSME scheme, which features both constant size ciphertexts and decryption keys via a deterministic key generation mechanism and a compact algorithm. In addition, the pre-setting of mapping for each element in attribute sets during the setup phase instead of encryption and keygen ones is also one of the reasons for our efficiency. Moreover, we prove the security of our scheme in the selective model under the Decisional Bilinear Diffie-Hellman Exponent assumption. Finally, we provide a theoretical comparison of performance and an experimental result, which shows that our scheme is really practical for privacy-preserving collaborative evaluation.