(全)同态加密在基于密文计算模型中的应用

Application of (Fully) Homomorphic Encryption for Encrypted Computing Models

随着云计算和大数据的快速发展,存储在云平台上数据的保密性、用户的隐私保护和数据商业利用一直是困扰学术界和产业界的难题.如何在保证数据机密性和用户隐私的基础上,实现数据的有效利用?(全)同态加密算法出现和快速发展为解决这个难题提供了一种有效方案选择.文章中首先介绍了云平台中存储数据安全性、用户隐私保护和数据商业利用这三者之间的关系和实现这三者之间平衡的重要性.接着,文章给出了基于理想格的Gentry原始方案、基于RLWE的BGV方案和FV方案的效率比较和分析.在上述的比较和分析中,文章给了具体的参数设置,详细地分析了秘钥生成时间、加解密时间和密文同态操作时间.通过分析发现,基于RLWE的somewhat同态加密方案为很多涉及实际问题的计算模型和算法给出了比较高效的解决方案.文章最后给出了基于BGV方案的两个somewhat同态加密应用案例,并且给出相关的效率分析.相关方案的效率分析和实验结果表明基于RLWE的somewhat同态加密方案是解决数据保密性、用户的隐私保护和数据商业利用的最有效方案.

With the rapid development of cloud computing and big data, data security, user privacy protection and commercial use of data in the cloud are challenging problems for academia and industry. How to ensure the effective use of data on the basis of ensuring data confidentiality and user privacy is an important problem.(Fully) homomorphic encryption provides one of the effective solutions to this problem. First of all, this paper introduces the relationship and balance among the data storage security, user privacy protection and business utilization of data in the cloud. Then,this paper presents efficiency comparison and analysis of the Gentry primitive scheme based on ideal lattice, BGV scheme and FV scheme from RLWE. Based on the comparison and analysis, this paper gives a specific parameter setting, a detailed analysis of the key generation time, decryption time and homomorphic evaluation time on encrypted data. It is found that the somewhat homomorphic encryption scheme based on RLWE provides a more efficient solution for many computational models and algorithms concerning practical problems. Finally, two cases based on BGV scheme and efficiency analysis are given. Efficiency analysis and experimental results show that the somewhat homomorphic encryption scheme from RLWE is the most effective solution for data privacy, user privacy protection and commercial utilization of data.