Securing cloud-based IoT applications with trustworthy sensing

Abstract

The omnipresence of resource-constrained sensors connected to the cloud has enabled numerous Internet of Things (IoT) applications. However, the trust in these IoT applications is severely compromised by security concerns. We introduce a lightweight and effective security approach for such applications by protecting the sensors. Our approach leverages Physically Unclonable Functions (PUF) on the sensor platform to ensure non-repudiation of sensed data and integrity of sensor hardware and firmware. We compare the performance of different PUF implementations on Atmel, ARM, and FPGA-based sensing platforms, analyze the security properties of the proposed protocols and determine the overhead in terms of latency, storage and logic area. Our evaluation shows that it only incurs insignificant overhead on low-end sensors.