Attribute-Based Content Redaction in Large-Scale Data Systems: A Study of Granular Access Control

Abstract

Data privacy has become significantly more important over the past years, leading to new laws and regulations that citizens and organizations must abide by. As a consequence, keeping data from being exposed to the wrong audience is no longer just an interest of the individual - it is also a legal requirement on companies that collect and store sensitive information. Different geographical regions may also enforce different data privacy laws, making matters even more complex for organizations that operate on a global scale. On top of the regulatory aspects, internal information security policies may specify that some subsets of data must be shared differently depending on security classifications and who requests it. This master’s thesis project conducts a Design Science Research study aiming to combine two existing techniques – attribute-based access control and redaction. The goal is to evaluate whether the resulting component is a viable approach to granular access control in request-response type APIs that expose sensitive data to a global audience. The study produces a Proof-of-Concept implementation as an artifact, which is evaluated and compared to the type of role-based RESTful APIs commonly used in industry today.