Middleware for privacy protection of ambient intelligence and pervasive systems

Ambient Intelligence (AmI) and pervasive environments pose security and privacy threats that reach way beyond the ones found in traditional systems. These threats arise from increased complexity and unpredictable nature of such systems. In such environments a significant amount of information directly related to users are fetched, exchanged, processed, inferred, aggregated and amplified throughout the system infrastructure. In order to solve such problems we present conceptual privacy enabling architecture of infrastructural middleware that uses systemic and manageable approach to enabling pirvacy using an orchestrated blend of trust management, privacy negotiation and identity management. Our work summarizes some of the results of the DAIDALOS IST FP6 Integrated Project[1] approach to enabling privacy. Entities involved in communication first have to negotiate privacy agreements based upon trustworthiness, afterwards appropriate virtual identities are created or selected and in the last phase privacy agreements violations are considered and trustworthiness of the communicating entities are updated accordingly. When inter-entity interaction is completed the experience gained in one communication cycle is fed back via updating of the reputation models. In this manner we approach to balancing the conflicting interest of AmI and pervasiveness on one hand and protecting privacy on the other.