Cloud computing has gained increasing acceptance in recent years. In privacy-conscious domains such as healthcare and banking, however, data security and privacy are the greatest obstacles to the widespread adoption of cloud computing technology. Despite enjoying the benefits brought by this innovative technology, users are concerned about losing the control of their own data in the outsourced environment. Encrypting data can resolve confidentiality and integrity challenges, but the key to mitigating users’ concerns and encouraging broader adoption of cloud computing is the establishment of a trustworthy relationship between cloud providers and users.
In this dissertation, we investigate a novel trust-oriented data protection framework adapted to the cloud environment. By investigating cloud data security, privacy, and control related issues, we propose a novel data protection approach that combines active and passive protection mechanisms. The active protection is used to secure data in an independent and smart data cube that can survive even when the host is in danger. The passive protection covers the actions and mechanisms taken to monitor and audit data based on third party security services such as access control services and audit services. Furthermore, by incorporating full mobility and replica management with the active and passive mechanisms, the proposed framework can satisfy confidentiality, integrity, availability, scalability, intrusion-tolerance, authentication, authorization, auditability, and accountability, increasing users’ confidence in consuming cloud-based data services.
In this work we begin by introducing cloud data storage characteristics and then analyse the reasons for issues of data security, privacy and control in cloud. On the basis of results of analysis, we identify desirable properties and objectives for protecting cloud data. In principle, cryptography-based and third party based approaches are insufficient to address users’ concerns and increase confidence in consuming cloud-based data services, because of possible intrusion attacks and direct tampering of data. Hence, we propose a novel way of securing data in an active data cube (ADCu) with smart and independent functionality. Each ADCu is a deployable data protection unit encapsulating sensitive data, networking, data manipulation, and security verification functions within a coherent data structure. A sealed and signed ADCu encloses dynamic information-flow tracking throughout the data cube that can precisely monitor the inner data and the derivatives. Any violations of policy or tampering with data would be compulsorily recorded and reported to bundled users via the mechanisms within the ADCu. This active and bundled architecture is designed to establish a trustworthy relationship between cloud and users.
Subsequently, to establish a more comprehensive security environment cooperating with an active data-centric (ADC) framework, we propose a cloud-based privacy-aware role-based access control (CPRBAC) service and an active auditing service (AAS). These components in the entire data protection framework contribute to the passive security mechanisms. They provide access control management and audit work based on a consistent security environment. We also discuss and implement full mobility management and data replica management related to the ADCu, which are regarded as significant factors to satisfy data accountability, availability, and scalability.
We conduct a set of practical experiments and security evaluation on a mini-private cloud platform. The outcome of this research demonstrates the efficiency, feasibility, dependability, and scalability of protecting outsourced data in cloud by using the trust-oriented protection framework. To that end, we introduce an application applying the components and mechanisms of the trust-oriented security framework to protecting eHealth data in cloud.
The novelty of this work lies in protecting cloud data in an ADCu that is not highly reliant on strong encryption schemes and third-party protection schemes. By proposing innovative structures, concepts, algorithms, and services, the major contribution of this thesis is that it helps cloud providers to deliver trust actively to cloud users, and encourages broader adoption of cloud-based solutions for data storage services in sensitive areas.