Weighted average problem revisited under hybrid and malicious model
Hu Xiong, Echo P. Zhang, Tat Wing Chim, Siu‐Ming Yiu, Lucas C. K. Hui · The HKU Scholars Hub (University of Hong Kong) · 2012
It is getting more common for two or more parties to jointly compute some statistics, say for marketing, by combining information on their private databases without disclosing the private data to the others. The core problem is usually known as secure multi-party computation (SMC). A number of solutions have been proposed. However, almost all of them assume a semi-honest model which is unrealistic. On the other hand, protocols that work under the malicious model (all participating parties can be malicious) are usually complicated and expensive in terms of communication and computation. In this paper, we try to consider a more reasonable model, the hybrid security model, in which at least one party is semi-honest. We want to make sure that the malicious parties will not get the correct final result if they perform malicious behaviors. We propose a scheme to solve the two-party weighted average problem (WAP) under this hybrid security model. We also show that the scheme can be extended to work under the malicious model using any fair exchange scheme. We formally show that our schemes are secure. We also implemented the schemes and showed that our scheme under the hybrid security model is reasonably fast and efficient for practical use.