Security for an Internet-of-Things Network in the Context of a Dairy Management System

Securing data in an Internet-of-Things network can be a challenge. Devices often operate on a very limited energy budget and thus common cryptographic solutions can quickly become too resource-hungry. This study provides a cryptographic system design focused on use in an Internet-of-Things network. The system design has been developed in the context of a precision dairy farming product, and runs on an EFR32MG22 microcontroller. Throughout this research we focus on identifying the performance gain that hardware acceleration of certain operations offers. We first perform an analysis of potential adversaries and the specific threats they pose. This was done using a combination of the STRIDE and DREAD frameworks for threat identification and ranking, as well as with a novel attacker profiling model that integrates with these frameworks. Next, we evaluate the performance of different symmetric encryption ciphers to determine which provides the most security whilst consuming as little energy as possible. We find that hardware-accelerated AES-CCM with key length of 256 bits consumes only roughly 4.5 mAh over a time span of ten years. To finalize our system design, we implement an identity-based key agreement protocol and evaluate its performance. This implementation shows that such a protocol is very suitable for this type of network, regardless of whether hardware acceleration is used or not.