节点文献
面向应用程序的数据安全外包存储协议的研究与实现
Research and Implementation of Application-Oriented Data Security Outsourcing Storage Protocol
【作者】 刘震;
【导师】 许春香;
【作者基本信息】 电子科技大学 , 网络空间安全, 2024, 硕士
【摘要】 随着计算机技术的快速发展,市场上出现了各种的应用程序。应用程序会在用户本地设备上产生大量数据,为存储能力有限的设备带来负担。为缓解存储负担,用户可将应用程序产生的数据外包到远程云服务器。安全的数据外包要求便携性和盲化性。便携性允许用户使用单一口令从不同的设备访问数据,而盲化性能够确保未被用户授权的实体无法获取数据内容。便携盲化云存储(Portable Blind Cloud Storage,PBCS)能够同时满足以上要求。然而,当应用服务器和云服务器同时被敌手妥协时,现有的PBCS方案容易遭受离线口令猜测攻击:敌手能利用妥协服务器获得的用户注册信息推断出用户的口令。而且,PBCS方案在修改文件时,需要将文件先请求回本地再修改,引入了大量通信开销。此外,PBCS方案不兼容密文去重,存储大量相同文件会浪费云服务器存储资源。因此,以PBCS为基础,本文开展面向应用程序的数据外包存储协议的研究设计,主要贡献如下:(1)为解决PBCS存在的安全问题,本文提出了抗妥协服务器的便携盲化云存储方案IPBCS。在注册阶段,IPBCS使用存储在可信执行环境中的密钥对用户口令进行强化。随后,用户使用强化口令生成认证令牌并将该令牌存储在服务器上作为注册信息。即使注册信息被敌手获得,在密钥受可信执行环境保护的条件下,IPBCS也能抵抗离线口令猜测攻击。此外,IPBCS采用了基于凭证/令牌的加密算法以支持在公开信道上的用户认证。安全分析和性能评估结果表明IPBCS能够确保用户数据的盲化性和完整性,且具有最优的计算、通信和存储开销。(2)为突破PBCS的功能局限,本文提出了支持文件修改和密文去重的便携盲化云存储方案De-PBCS。De-PBCS将文件分割成多个文件块,使用同态可聚合向量承诺算法产生一个用于验证文件完整性的承诺以及每个文件块的成员证明。同态性允许用户在修改文件块时,同步更新承诺和成员证明而不用将整个文件请求回本地。可聚合性支持用户对文件块的批量查询和验证。同时,为了兼容密文去重,De-PBCS将应用服务器作为辅助服务器,使用服务器辅助的去重机制拓展了PBCS。安全分析和性能评估结果表明De-PBCS能够保证用户数据安全并具有良好的计算、通信和存储开销。
【Abstract】 With the rapid development of computer technology,various applications(Apps)have emerged in the market.These Apps generate massive data on users’ storage-limited devices,which burdens them.To alleviate this burden,users can outsource the Appgenerated data to a remote cloud server.Secure data outsourcing requires portability and blindness.Portability allows users to access data from different devices using a single password,while blindness ensures that unauthorized entities cannot access the content of the data.Portable Blind Cloud Storage(PBCS)can meet these requirements simultaneously.However,existing PBCS schemes are vulnerable to offline password guessing attacks(OPGA)when both the App server and cloud server are compromised by adversaries.Adversaries can use registered user information they obtain by compromising servers to defer the user’s password.Additionally,PBCS schemes introduce significant communication overheads when modifying files,since the file needs to be requested back to the local device before modification.Furthermore,PBCS schemes are incompatible with ciphertext deduplication,which wastes cloud server storage resources when storing a large number of identical files.Therefore,based on PBCS,this thesis conducts research and design on App-oriented data outsourcing storage protocol,with the following contributions:(1)To address the security issues of PBCS,this thesis proposes an Improved Portable Blind Cloud Storage scheme(IPBCS)that resists compromised servers.In the registration phase,IPBCS hardens the user’s password using a key stored in trusted execution environments(TEE).Subsequently,users generate an authentication token using the hardened password and store this token on the server as registration information.Even if the registration information is obtained by adversaries,IPBCS can resist OPGA as long as the key remain protected in TEE.Furthermore,IPBCS adopts a credential/token-based encryption algorithm to support user authentication over public channels.Security analysis and performance evaluation results demonstrate that IPBCS can ensure the blindness and integrity of data,while having optimal computational,communication,and storage costs.(2)To overcome the functional limitations of PBCS,this thesis proposes a scheme called De-PBCS that supports file modification and ciphertext deduplication.De-PBCS divides files into multiple file blocks and uses a homomorphic aggregatable vector commitment algorithm to generate a commitment for verifying the integrity of the file,as well as membership proofs for each file block.Homomorphism allows users to synchronously update commitments and membership proofs when modifying file blocks without requesting the entire file back to the local device.Aggregatability supports batch queries and verification of file blocks by users.Meanwhile,to be compatible with ciphertext deduplication,De-PBCS extends PBCS by utilizing the App server as an auxiliary server and employing a server-aided deduplication mechanism.Security analysis and performance evaluation results show that De-PBCS can guarantee the security of user data and has good computation,communication,and storage costs.
【Key words】 Password Authentication; Password Guessing Attacks; Ciphertext Deduplication; Homomorphic Updatable Encryption; Vector Commitment;
- 【网络出版投稿人】 电子科技大学 【网络出版年期】2025年 07期
- 【分类号】TP309.2;TP333