In a context-aware service platform, service providers adapt their services to the current situation of the service users using context information retrieved from context information providers. In such a service provisioning platform, important trust and privacy issues arise, because different entities responsible for different tasks have to collaborate in the provisioning of the services. Context information is privacy sensitive by nature, making the communication and processing of this information a potential privacy threat. User acceptance of context-aware services depends on the users' perception of how the context-aware service platforms deal with their privacy. Users of context-aware services need to control who is authorized to access their context information, and how their context information is communicated and processed after the access is granted. Providing users with control over their privacy is especially difficult in context-aware service platforms, since users' privacy desires are personalized to the context situation these users are in. Users have, for example, different wishes regarding privacy of their health data when they are being treated in a hospital than when they are in their working environment. For users to feel in control of their privacy, the mere specification of their privacy preferences is not enough in the trade-off between privacy and context-aware service adaptation. Users must also be confident that the specified privacy preferences are being enforced by the entities of the service platform that are responsible for communication and processing their context information, such as the context information providers. Trust, therefore, is an integral part of the users' privacy concerns in context-aware service platforms. In the trade-off we address in this thesis, context-aware service providers are more concerned with their capability of providing reliable context-based service adaptation because this is their primary business goal. To be able to reliably adapt, service providers depend on the trustworthiness of the context information providers that provide the context information about the service users. Privacy issues are also important for service providers because the reporting in the media of privacy incidents involving their service provisioning infrastructure also impact their primary business due to the loss of reputation. Existing trust and privacy solutions targeted at context-aware service platforms fail to address the different trust aspects and dependencies between the entities participating in a context-aware service provisioning platform. Existing solutions focus on at most one trust aspect at a time, for example, privacy enforcement or identity certification, and do not consider dependencies between the different aspects that are present in the trade-off we address in this thesis. Other concerns of users and service providers such as reliability of the context-aware service adaptation, or the relationship between quality aspects of the context information and trust are not addressed by existing solutions in an integrated way. Furthermore, existing trust and privacy management solutions for context-aware service platforms offer poor support for personalized context-based privacy management. In this thesis we present the analysis, design, implementation, and evaluation of a trust and privacy preferences management solution to support service users and service providers of a context-aware service platform. The functionality of this solution consists of three major contributions that focus on trust and privacy issues from the perspective of users and service providers. The first major contribution of this thesis is a trust-based selection mechanism that support users of context-aware services in selecting trustworthy service providers to interact with. This mechanism supports the users in this selection process, taking into account the users' goals, trust beliefs, and the trust dependencies between the service users and the entities that collaborate in the context-aware service provisioning. The service users' goals we use as input to our mechanism are related to the trade-off between privacy protection and context-aware service adaptation. The second major contribution of this thesis is a trust-based selection mechanism that support context-aware service providers in selecting trustworthy context information providers. This mechanism supports service providers in selecting context information providers taking into account their trustworthiness to provide context information about a specific user and quality level. This mechanisms contributes to the improvement of the context-based adaptation capabilities of context-aware service providers. The third major contribution of this thesis is a generic context-based policy management concept called a Context-Aware Management Domain (CAMD). The CAMD concept is used by us in our case studies to support the specification of trust and privacy policies by means of context-based authorizations and obligations. Our CAMD concept is realized using policy-based management, and uses context information as input for the policy management task. The objective of our CAMD concept is to support users and system administrators in managing policies aimed at controlling who is authorized to access the users' context information at what quality level, and which actions these entities are obliged to execute after access is granted. One example of a privacy obligation is to delete all location information about a user after a context condition is satisfied. We have evaluated the technical feasibility of our contributions through case studies and prototype implementations. We have also evaluated the usability and usefulness aspects of our contributions from a user perspective through an user survey. Our technical and user survey evaluations show that our solutions are technically feasible and that the majority of the survey participants were able to understand and believe that our contributions are useful. Furthermore, our technical feasibility and user survey evaluations contribute to increased knowledge about the trust and privacy requirements of a context-aware service platform with examples of context-based policies and user goals when using a context-aware service.
|Award date||30 Mar 2012|
|Place of Publication||Enschede|
|Publication status||Published - 30 Mar 2012|
- Context aware
- Trust management
- Privacy management