Computation offloading and energy consumption in the Internet of Things: a structured approach

Internet of Things (IoT) and wearable devices often have limited energy budgets and reduced computing resources. Continuously running an application on a smartwatch, for example, can quickly drain its battery life, resulting in an unsatisfactory user experience. Computation offloading, i.e. transferring computationally intensive tasks to more powerful devices in the edge-cloud continuum, can help reduce the power needs of IoT and wearable devices, possibly extending their battery life. In this thesis, the application running on resource-constrained devices is structured as a set of functions that can be relocated on the edge according to a serverless approach. The ultimate goal is to determine whether this paradigm can be useful in optimizing a wearable device's battery life and performance.