Modeling self-adaptative IoT architectures

Abstract

An Internet of Things (IoT) architecture describes a network of (physical) objects exchanging data over the Internet. In complex IoT deployments, as those typical of Industry 4.0 systems, this architecture comprises four different layers: device, edge, fog, and cloud. Moreover, the composition and computation tasks assigned to each layer may dynamically evolve due to environmental changes.As such, modeling IoT architectures is a complex process that must cover as well the specification of self-adaptation rules to ensure the optimized execution of the IoT system. To facilitate this task, we propose a new IoT Domain-Specific Language (DSL) covering both the static and dynamic aspects of an IoT deployment. The definition of the DSL, its projectional-based editor and a first prototype of a Kubernetes manifest generator are available as open source software.