An Infrastructure for Privacy-Enabled Sharing of Context Data

Hynes, Gearoid
The concept of context originated in the ubiquitous computing domain where information regarding the user's situation was gathered using sensing devices, allowing the application to automatically adapt according to the user's current situation. Context awareness enabled the development of smart spaces in a multitude of different environments such as offices, homes, cars, hospitals and universities. Traditionally, the smart space that produced the context data was the sole consumer of that data and accessing the context data from third-party applications was cumbersome. Consequently, smart spaces became stovepipes; having little integration beyond their isolated application scenarios. The development of context-aware applications was limited to environments with access to rich sources of context data and which had large resources available for the development of context-aware applications. The development of these context-aware applications required expertise in the areas of context modelling, context gathering, context management and data privacy. As a result, the inclusion of context-aware behaviours in applications outside of ubiquitous computing research prototypes has been limited. This thesis defines and presents a distributed web-based model for the controlled shar-ing of context data between heterogeneous third-parties. A web-based approach is adopted in order to maximize the variety of potential applications, such as mobile, desktop, smart space and web-based applications. This `Context as a Service' (CxaaS) approach to context management enables applications to access context-data from domains beyond their usual operational boundaries. The support for ubiquitous computing scenario oriented sharing of context data has been addressed in three key areas: (i) discovery of context data from remote sources concerning new and returning users, (ii) enabling the privacy sensitive sharing of context data between heterogeneous third-parties, and (iii) provision of a formal context model appropriate for use with the discovery and privacy mechanisms. The model defined in this thesis is evaluated with an implementation of a comprehensive business use-case. A RESTful approach is adopted in this particular instantiation of the CxaaS model which facilitates the rapid development of context-aware applications through the use of popular libraries. The core CxaaS context ontology is supported and mechanisms are provided to extend this model if necessary.
Publisher DOI
Attribution-NonCommercial-NoDerivs 3.0 Ireland