一种基于CPU-GPU混合系统的并行同态加密算法

A Parallel Homomorphic Encryption Based on CPU-GPU Hybrid System

随着云计算服务模式的推广,用户数据的私密性逐渐成为云安全体系需要重点关注的焦点。在信息安全领域,数据加密是避免信息遭受非法访问的有效途径。同态加密是一种可直接对密文进行运算的加密算法,在给云服务提供商省去解密的复杂操作同时,也很好地保护了数据的私密性。但是,由于同态加密算法的计算复杂度特别高,导致给计算资源增加了繁重的负担。该研究针对上述问题设计了一种快速的并行计算框架,提出一种基于CPU-GPU混合系统的并行同态加密算法。文章提出的同态加密并行计算框架,主要致力于利用GPU来加速同态加密算法的运行,从而缩短同态加密算法的运行时间,保证云计算环境用户数据的安全。另外,论文设计了一种数据处理的流水线模式,从而进一步提高并行同态加密算法的并行度。实验也表明,论文提出的并行同态加密算法相比串行同态加密算法,在执行同态乘法和同态加法操作时,在运算时间上分别提高了70%和91%。

With the popularization of cloud computing service mode,the privacy of data has gradually become the focus of cloud security.In the field of information security,data encryption is an effective way to avoid illegal access to information.Homomorphic encryption is a kind of encryption algorithm that can directly operate on the ciphertext.Therefore,while saving the complex operation of decryption for cloud service providers,it also protects the privacy of data.However,due to the high computational complexity of the homomorphic encryption algorithm,it adds heavy workload to the computing resources.In this paper,a fast parallel computing framework is designed for the above problems,and a parallel homomorphic encryption algorithm based on CPU-GPU hybrid system is proposed.This paper proposes a parallel computing framework for homomorphic encryption,which is mainly dedicated to using GPU to accelerate the operation of the homomorphic encryption algorithm.In special,the parallel computing scheme can shorten the operation time of the homomorphic encryption algorithm and ensure the security of user data in the cloud computing environment.This paper also makes further efforts to design a pipeline architecture of processing stream in CPU-GPU hybrid system to accelerate encryption.The experiment results show that this method gains more than 70%improvement(run time)and 91%improvement compared to the serial homomorphic multiplication and homomorphic addition respectively.