A GENERIC MULTI-TIER PRIVACY MODEL PREFERENCE SELECTION (GM-PMPS) IN A PERVASIVE ENVIRONMENT

Abstract

Pervasive computing is an emerging computing paradigm expected to become part of our everyday lifestyle in the foreseeable future. Despite its dynamic nature and high demand for information, many drawbacks and undesirable use in terms of privacy can be foreseen. More precisely, the pervasive computing paradigm raises concerns about end-user privacy, and ensuring privacy is becoming a major challenge requiring a tradeoff between privacy and context-aware service adaptation. This research work proposes a generic multitier model for end-user privacy preference selection to handle possible malicious requests through a predefined "aura" configured and controlled by users via privacy preferences. The multitier model is structured around users’ natural relations, categorized as personal, social, and third-party aura, which can be evaluated in a group for any privacy-related requests based on trust accumulated through formulated and archived reputations. Since the exchange of local trust is the basis for determining reputation, the necessary trust value is determined by the weighted average result of a reputation figure gathered from direct and indirect request responses of nodes within the established aura. Finally, the implemented prototype of the proposed model determines the trust level of the requesting node based on the user’s privacy preference selection bias point for the service and decides whether to respond automatically, require manual intervention, or block the request.