PARALLEL DATA LAB 

PDL Abstract

Poster Abstract: BUFS: Towards Bottom-Up Foundational Security for Software in the Internet-of-Things

1st IEEE/ACM Symposium on Edge Computing (SEC 2016), October 2016.

Jiaqi Tan, Rajeev Gandhi, Priya Narasimhan

Carnegie Mellon University

http://www.pdl.cmu.edu/

The Internet-of-Things (IoT) is a rapidly growing phenomenon. While IoT-enabled objects can provide rich features that can improve users’ lives, security failures can lead to severe consequences, particularly in safety-critical domains such as medical devices and automobiles. In addition, IoT-enabled objects are often connected to the Internet, increasing their risk for external attacks. Thus, it is important for IoT systems to have strong security guarantees. Some of the security challenges IoT systems face include the need for lightweight cryptographic algorithms and secure communications protocols. In practice, security mechanisms are implemented in a software stack on IoT devices. This software stack needs to (i) provide security mechanisms correctly, and (ii) faithfully execute application logic, without being circumvented by attackers. Software vulnerabilities may allow external attackers to circumvent these security measures: over 250 vulnerabilities were discovered in the top 10 IoT devices in use today in a recent study [1]. We propose BUFS, a bottom-up and foundational approach for verifying the security of the software stack in an IoT system, to provide guarantees for how the software is secure.

FULL PAPER: pdf