Scalability Evaluation of a Per-User Access Control Framework

Today's Internet of Things (IoT) devices have a variety of security requirements and policies. While an access control is applied to such devices to meet the varieties of requirements and policies, the access control has rarely been used for network resources. Due to this situation, we have proposed a per-user access control framework, which realizes the access control for network links and bandwidth as network resources by using Software-Defined Networking, in our previous work. The proposed framework enables a network administrator to apply access control to network resources simply by giving the administrator's policy as input to the proposed framework. However, there remains the concern that the proposed framework may cause a significant overhead for the data transfers when the number of IoT devices is increased. In this paper, we investigate how scalable the proposed framework is as infrastructure, by considering the actual and practical situation where lots of IoT devices are used. Our evaluation results imply that the overhead incurred by the proposed method is negligible, especially in the case where IoT devices transfer large-sized data. Also, the evaluation results show that the proposed framework reduces the exposure time of the IoT devices to a third party.