On information-centric resiliency and system-level security in constrained, wireless communication

Abstract

The Internet of Things (IoT) interconnects many heterogeneous embedded devices either locally between each other, or globally with the Internet. These things are resource-constrained, e.g., powered by battery, and typically communicate via low-power and lossy wireless links. Communication needs to be secured and relies on crypto-operations that are often resource intensive and in conflict with the device constraints. These challenging operational conditions on the cheapest hardware possible, the unreliable wireless transmission, and the need for protection against common threats of the inter-network, impose severe challenges to IoT networks.

In this thesis, we advance the current state of the art in two dimensions. Part I assesses Information-centric networking (ICN) for the IoT, a network paradigm that promises enhanced reliability for data retrieval in constrained edge networks. ICN lacks a lower layer definition, which, however, is the key to enable device sleep cycles and exclusive wireless media access. This part of the thesis designs and evaluates an effective media access strategy for ICN to reduce the energy consumption and wireless interference on constrained IoT nodes. Part II examines the performance of hardware and software crypto-operations, executed on off-the-shelf IoT platforms. A novel system design enables the accessibility and auto-configuration of crypto-hardware through an operating system. One main focus is the generation of random numbers in the IoT. This part of the thesis further designs and evaluates Physical Unclonable Functions (PUFs) to provide novel randomness sources that generate highly unpredictable secrets, on low-cost devices that lack hardware-based security features. This thesis takes a practical view on the constrained IoT and is accompanied by real-world implementations and measurements. We contribute open source software, automation tools, a simulator, and reproducible measurement results from real IoT deployments using off-the-shelf hardware. The large-scale experiments in an open access testbed provide a direct starting point for future research.