MATS: A Multi-aspect and Adaptive Trust-based Situation-aware Access Control Framework for Federated Data-as-a-Service Systems

Abstract

Federated Data-as-a-Service systems are helpful in applications that require dynamic coordination of multiple organizations, such as maritime search and rescue, disaster relief, or contact tracing of an infectious disease. In such systems it is often the case that users cannot be wholly trusted, and access control conditions need to take the level of trust into account. Most existing work on trust-based access control in web services focuses on a single aspect of trust, like user credentials, but trust often has multiple aspects such as users’ behavior and their organization. In addition, most existing solutions use a fixed threshold to determine whether a user’s trust is sufficient, ignoring the dynamic situation where the trade-off between benefits and risks of granting access should be considered. We have developed a Multi-aspect and Adaptive Trust-based Situation-aware Access Control Framework we call “MATS” for federated data sharing systems. Our framework is built using Semantic Web technologies and uses game theory to adjust a system’s access decisions based on dynamic situations. We use query rewriting to implement this framework and optimize the system’s performance by carefully balancing efficiency and simplicity. In this paper we present this framework in detail, including experimental results that validate the feasibility of our approach.