Content area
This dissertation aims to perform a comprehensive risk analysis of the possible threats presented to a user while trying to sign on to a service that uses the OpenID Connect (OIDC) protocol for authentication. This analysis counts with the help of fuzzy logic models, integrating proofs of privacy compliance to ensure conformity with GDPR (General Data Protection Regulation).Research was conducted on various concepts related to this theme, focussing on the OIDC authentication mechanism and the potential risks associated with user privacy consent when accessing a service through the authentication process. The concepts of fuzzy logic and fuzzy logic models, also known as fuzzy inference systems, were explored in-depth due to their advantage of being ideal for risk analysis. Within fuzzy inference systems, the focus was specifically on three types: Mamdani, Sugeno, and Tsukamoto, which were then compared to determine which would be the most suitable for the context of this dissertation.The development of a risk module is also documented, as it was the main tool for performing the risk analysis. It comprises different components, such as two Application Programming Interfaces (APIs), one database, and a fuzzy risk system. This is the main component of the module since it is responsible for performing the calculations of the risk related to the claims (user information) that the service tries to retrieve during the user-authentication process. Before being considered complete, this fuzzy system suffered many iterations, to improve its effectiveness in the final risk calculations.The risk module, which underwent a performance test, is also responsible for providing GDPR compliance and risk of the service to the end-user interface, which presents this information to the user who is trying to authenticate. This end-user interface design was carried out using a user-centric approach.The dissertation confirmed the success of the risk assessment using fuzzy logic models in a service that uses authentication through the OIDC protocol, highlighting that fuzzy logic is appropriate for risk analysis. In terms of end-user evaluation, it was also favourable, as most of the participants confirmed that they experienced greater safety awareness with the developed solution.