【作者】 刘燕；

【导师】 刘振华；

【作者基本信息】 西安电子科技大学 ， 应用数学， 2020， 硕士

【摘要】 网络技术的不断进步,使云计算得到了广泛应用.云存储作为云计算的一部分,以其价格低廉、按需付费的优势能为数据用户带来了便利.数据直接以明文形式上传存储在云服务器中,必然会泄露数据的机密性,因此需要先进行加密处理.然而,加密会阻碍数据搜索.为解决密文数据的关键词搜索问题,Song提出了可搜索加密概念.由于传统的可搜索加密方案无法提供灵活的数据共享,学者们融合属性基加密的细粒度访问控制优点,设计了属性基可搜索加密算法.近年来,属性基可搜索加密算法取得了重大进展,但仍存在一些亟待解决的问题:(1)数据的完整性验证问题——云服务器是半可信的,可能会为了节省资源只存储部分数据,或者在搜索阶段只返回部分数据;(2)属性撤销问题——用户私钥可能丢失或被盗,需要撤销或更新其解密权限;(3)访问结构的隐私保护问题——加密数据的访问结构会泄露用户的隐私;(4)恶意用户追踪问题——多个用户可能拥有相同的解密权限,某些用户可能会为了经济利益泄露密钥而不会被追责.这些问题制约着属性基可搜索加密算法的实际应用.为解决上述问题,本文主要研究工作如下:(1)提出了一种具有验证和属性撤销功能的属性基可搜索加密方案.针对电子医疗记录系统中云服务器不可信的问题,可信第三方审计采用聚合签名技术与云服务器进行交互,能够验证数据的完整性;针对系统中医生职位变更引起的属性撤销问题,采用密钥加密树和最小子集覆盖方法管理属性组密钥,实现了属性撤销.最后,利用可证明安全理论对方案的安全性进行了证明.同时,性能分析表明方案具有一定的实用性.(2)提出了一种支持隐私保护和用户追踪的属性基可搜索加密方案.考虑到属性基可搜索加密中访问结构会在一定程度上泄露用户隐私,该方案利用AND门多值独立访问结构进行加密,密文数据中不包含访问结构,从而可以有效防止敏感隐私信息泄露.为解决恶意用户追踪问题,该方案采用Shamir门限技术,实现了恶意用户的追踪功能,并追踪存储消耗为常量级.最后,方案的安全性证明和性能分析分别表明方案是安全的和高效的.更多还原

【Abstract】 With the continuous improvement of network technology,cloud computing has been more and more widely applied.As a part of cloud computing,cloud storage can bring convenience to data users with advantages of low price and pay-as-you-go.If the data are uploaded directly and stored in the cloud server in the form of plaintext,then it is inevitably to disclose the confidentiality of the data,and thus it is necessary to encrypt-then-upload the data.However,encryption will hinder the data retrieval.To solve the problem of keyword search over encrypted data,Song proposed the concept of searchable encryption.Since traditional searchable encryption schemes cannot provide the flexible data sharing,some scholars have combined them with attribute-based encryption due to its advantages of fine-grained access control to propose attribute-based searchable encryption schemes.In recent years,attribute-based searchable encryption schemes have been made significant progress,but there are still some problems that need to be solved:(1)data integrity verification—the cloud server may be semi-trusted,and only store a part of data in order to save resources,or only return a part of the data during the search phase;(2)attribute revocation—the users’ private keys may be lost or stolen,and thus their decryption rights needs to be revoked or updated;(3)privacy preserving of access policy—the access structure attached in the encrypted data will reveal the users’ privacy;(4)malicious user trace—many users may have the same decryption rights for some ciphertext,but some users may disclose the key for financial benefit without being accountable.These problems will limit the applications of attribute-based searchable encryption schemes.In order to solve the above problems,the main efforts of this paper are as follows:(1)To propose a verifiable attribute-based keyword search encryption scheme with attribute revocation.For the problem that the cloud server is not trusted in the electronic medical record system,a trusted third-party auditor utilizes the aggregated signature technology to interact with the cloud server,which can verify the data integrity.Considering the attribute revocation problem caused by the change of the doctor’s position in the system,the key encryption tree and the minimum subset coverage method are used to manage the attribute group keys,which can realize attribute revocation.Finally,the propose scheme is proved to be secure by the use of security reduction,and the experimental simulations show that the proposed scheme is practical.(2)To present a privacy-preserving attribute-based encryption scheme with keyword search and user tracing.Considering that the access structure in attribute-based searchable encryption will leak the users’ privacy to a certain extent,the proposed scheme utilizes AND gate multi-value independent access structure to encrypt the data,and the access structure is not attached in the ciphertext,which can effectively prevent the sensitive privacy information from leaking.To solve the problem of malicious user tracing,the proposed scheme uses Shamir’s threshold technology to trace malicious users,and the tracing storage consumption is constant.Finally,the detailed security proofs and the performance analysis show the proposed scheme is secure and efficient.更多还原