In this paper, we present a critical assessment of the use of device fingerprinting for risk-based authentication in a state-of-practice identity and access management system. Risk-based authentication automatically elevates the level of authentication whenever a particular risk threshold is exceeded. Contemporary identity and access management systems frequently leverage browser-based device fingerprints to recognize trusted devices of a certain individual. We analyzed the variability and the predictability of mobile device fingerprints. Our research shows that particularly for mobile devices the fingerprints carry a lot of similarity, even across models and brands, making them less reliable for risk assessment and step-up authentication.
展开▼