Access Control Models and Technologies for the Internet of Things

点击翻译按钮获取中文摘要

Access Control is the security feature of a computer system that ensures access to resources is only granted to authorized users under legitimate circumstances. Ross Anderson once said that Access Control is the "traditional center of gravity of computer security." The advent of the Internet of Things (IoT), namely, the inter-connected network of smart devices that are embedded in our physical environment, necessitates a significant rethinking of how Access Control is to be achieved. Mindlessly deploying IoT devices into our smart homes, vehicular networks, and industrial control systems (ICS) without properly addressing these emerging challenges will not only cause us loss of valuable digital assets, but also impact our physical safety. In this work, I will address three access control challenges in current IoT environments. The first challenge arises from the sheer scale of the massively distributed system composed of inter-connection of smart devices. How does one write down a properly formulated access control policy that accounts for the many devices and unanticipated access scenarios? Note that any errors in policy formulation could lead to the leakage of access to devices that affects our safety. With IoT becoming a global phenomenon, it is unrealistic to assume that all the users and devices are registered to a single Identity Management System. That assumption is as unreasonable as asking everyone in the world to have a Facebook or Google account (or to use the same Internet Service Provider). The second protection challenge is to support scenarios in which users (or their devices) who do not know one another want to share resources when they encounter one another in a casual and transient manner. The crux is to enable total strangers to establish trust of one another during chance encounters, without resorting to a global Identity Management System. The third protection challenge concerns the fact that smart devices are embedded in the physical world. Misuse is not simply about whether a single access has been granted to unauthorized parties, but about how, over time, the access pattern that emerge. The challenge is about the control of behavior pattern, as well as the pattern of collaboration among multiple devices. The objective of the proposed research program is to design, analyze, and develop novel Access Control models and technologies that address the aforementioned challenges.