Establishing Confidence Level Measurements for Remote User Authentication in Privacy-Critical Systems

Abstract

User Authentication is the process of establishing confidence in the User identities presented to an information system. This thesis establishes a method of assigning a confidence level to the output of a user authentication process based on what attacks and threats it is vulnerable to. Additionally, this thesis describes the results of an analysis where the method was performed on several different authentication systems and the confidence level in the authentication process of these systems determined. Final conclusions found that most systems lack confidence in their ability to authenticate users as the systems were unable to operate in the face of compromised authenticating information. Final recommendations were to improve on this inadequacy, and thus improve the confidence in the output of the authentication process, through the verification of both static and dynamic attributes of authenticating information. A system that operates confidently in the face of compromised authenticating information that utilizes voice verification is described demonstrating the ability of an authentication system to have complete confidence in its ability to authenticate a user through submitted data.