Towards Efficient Evaluation of ABAC Policies using High-Dimensional Indexing Techniques
Proteet Paul, Shamik Sural · 2021
Access mediation in Attribute-Based Access Control (ABAC) is done through the evaluation of a set of rules comprising the organizational security policy. There is often a need to relax the exact matching of rules while deciding on an access request, a typical example of which is emergency access to medical data of patients. However, such relaxations should only be done in a controlled manner and at the same time, they should be computationally efficient. In this paper, we show a direction towards effective rule similarity search in ABAC using techniques for indexing high-dimensional databases and associated queries. Since ABAC attributes can take both continuous as well as discrete values, we consider using C-ND Tree (Continuous and Non-Ordered Discrete Tree), which was originally developed for indexing high-dimensional data in hybrid spaces. Our preliminary experiments show encouraging results.