The LoRaWAN 1.0 has been shown to suffer from security vulnerabilities which require updating the LoRaWAN implementation on respective sensor nodes. However, updating the firmware of LoRaWAN end devices is a demanding task, as data rate and duty cycle limit the throughput to only a few kilobytes per second. Heterogeneity amongst sensors exacerbates the situation by requiring dedicated images for each sensor type. The thesis addresses both problems by improving the updatability of LoRaWAN end devices by allowing to replace single functions of the LoRaWAN stack with architecture-independent drop-ins based on WebAssembly.

KuVS Best Thesis Award 2020