Section: Application Domains

Application Domains

data privacy, personal information management, healthcare, ambient intelligence

Our work addresses varied application domains. Typically, data management techniques on chip are required each time data-driven applications have to be embedded in ultra-light computing devices. This situation occurs for example in healthcare applications where medical folders are embedded into smart tokens (e.g., smart cards, secured USB keys), in telephony applications where personal data (address book, agenda, etc.) is embedded into cellular phones, in sensor networks where sensors log row measurements and perform local computation on them, in smart-home applications where a collection of smart appliances gather information about the occupants to provide them a personalized service, and more generally in most applications related to ambient intelligence.

Safeguarding data confidentiality has become a primary concern for citizens, administrations and companies, broadening the application domains of our work on access control policies definition and enforcement. The threat on data confidentiality is manifold: external and internal attacks on the data at rest, on the data on transit, on the data hosted in untrusted environments (e.g., Database Service Providers, Web-hosting companies) and subject to illegal usage, insidious gathering of personal data in an ambient intelligence surrounding. Hence, new access control models and security mechanisms are required to accurately declare and safely control who is granted access to which data and for which purpose.

While the application domain mentioned above is rather large, one application is today more specifically targeted by the SMIS project. This application deals with privacy preservation in EHR (Electronic Health Record) systems. Several countries (including France) launched recently ambitious EHR programs where medical folders will be centralized and potentially hosted by private Database Service Providers. Centralization and hosting increase the risk of privacy violation. In 2007, we launched two projects (PlugDB and DMSP) tackling precisely this issue, with the final objective to experiment our technologies in the field. In 2011, we launched a new project (KISS) capitalizing on the previous ones and extending their scope towards the protection of any personal data delivered to individuals in an electronic form.