Serverless Edge Computing: Resource Scheduling Algorithms and Software Systems

Deploying AI-driven Internet of Things (IoT) applications in remote areas, such as farms and forests, poses significant challenges due to issues like limited internet access and electricity infrastructure constraints. This thesis proposes serverless edge computing to simplify the deployment of AI-driven IoT applications on devices powered by batteries and environmentally harvested energy sources like solar power. The results demonstrate remarkable performance improvements of up to 40%, primarily achieved through software enhancements.