In a database-as-a-service(DaaS)model,a data owner stores data in a database server of a service provider,and the DaaS adopts the encryption for data privacy and indexing for data query.However,an attacker can obtain ...In a database-as-a-service(DaaS)model,a data owner stores data in a database server of a service provider,and the DaaS adopts the encryption for data privacy and indexing for data query.However,an attacker can obtain original data’s statistical information and distribution via the indexing distribution from the database of the service provider.In this work,a novel indexing schema is proposed to satisfy privacy-preserved data management requirements,in which an attacker cannot obtain data source distribution or statistic information from the index.The approach includes 2 parts:the Hash-based indexing for encrypted data and correctness verification for range queries.The evaluation results demonstrate that the approach can hide statistical information of encrypted data distribution while can also obtain correct answers for range queries.Meanwhile,the approach can achieve nearly 10 times and 35 times improvement on encrypted data publishing and indexing respectively,compared with the start-of-the-art method order-preserving Hash-based function(OPHF).展开更多
数据库即服务(database as a service,DaaS)作为一种新型的数据存储提供模式被广泛应用.随着大数据时代的到来,数据量急剧增加,DaaS模式下的数据布局问题显得更加重要,即服务提供商如何根据应用中不同数据的性能需求对数据进行合理布局...数据库即服务(database as a service,DaaS)作为一种新型的数据存储提供模式被广泛应用.随着大数据时代的到来,数据量急剧增加,DaaS模式下的数据布局问题显得更加重要,即服务提供商如何根据应用中不同数据的性能需求对数据进行合理布局,将会对提高服务质量、增强用户体验和降低自身服务成本产生重要影响.然而对于服务提供者来说提高服务质量和降低服务成本是一对矛盾的目标.提出DaaS模式下的数据布局图概念,应用Pareto最优思想适合于解决多目标矛盾性问题的特点,给出一个基于性能-代价均衡的多节点DaaS数据布局策略.通过与随机策略和贪婪策略等传统策略的实验比较,方法能保证DaaS服务提供商用尽可能少的代价为用户提供更好的服务质量,实现服务质量与资源代价两个目标的均衡.展开更多
基金the National Natural Science Foundation of China(No.61931019).
文摘In a database-as-a-service(DaaS)model,a data owner stores data in a database server of a service provider,and the DaaS adopts the encryption for data privacy and indexing for data query.However,an attacker can obtain original data’s statistical information and distribution via the indexing distribution from the database of the service provider.In this work,a novel indexing schema is proposed to satisfy privacy-preserved data management requirements,in which an attacker cannot obtain data source distribution or statistic information from the index.The approach includes 2 parts:the Hash-based indexing for encrypted data and correctness verification for range queries.The evaluation results demonstrate that the approach can hide statistical information of encrypted data distribution while can also obtain correct answers for range queries.Meanwhile,the approach can achieve nearly 10 times and 35 times improvement on encrypted data publishing and indexing respectively,compared with the start-of-the-art method order-preserving Hash-based function(OPHF).
文摘数据库即服务(database as a service,DaaS)作为一种新型的数据存储提供模式被广泛应用.随着大数据时代的到来,数据量急剧增加,DaaS模式下的数据布局问题显得更加重要,即服务提供商如何根据应用中不同数据的性能需求对数据进行合理布局,将会对提高服务质量、增强用户体验和降低自身服务成本产生重要影响.然而对于服务提供者来说提高服务质量和降低服务成本是一对矛盾的目标.提出DaaS模式下的数据布局图概念,应用Pareto最优思想适合于解决多目标矛盾性问题的特点,给出一个基于性能-代价均衡的多节点DaaS数据布局策略.通过与随机策略和贪婪策略等传统策略的实验比较,方法能保证DaaS服务提供商用尽可能少的代价为用户提供更好的服务质量,实现服务质量与资源代价两个目标的均衡.
