Type, pad, and place: Avoiding data leaks in Cloud-IoT FaaS orchestrations

Placing applications composed as orchestrated serverless functions onto Cloud-loT infrastructures is a chal-lenging problem as it must consider hardware, software, network Quality of Service, and service interactions constraints. In this paper, we propose a novel declarative methodology that handles all of the above, also relying on information-flow analyses and padding techniques to prevent information leaks through side channels. A motivating use case from augmented reality is used to showcase the open-source prototype implementing our proposal.