DingNet: A Self-Adaptive Internet-of-Things Exemplar
by Michiel Provoost and Danny Weyns
Abstract
Recent efforts have shown that research on self-adaptive systems can benefit from exemplars to evaluate and compare new methods, techniques and tools. One highly relevant application domain for self-adaptation is the Internet-of-Things (IoT). While some initial exemplars have been proposed for IoT, these exemplars are limited in scope to support research in realistic IoT domains, such as smart cities. To address this limitation, we introduce the DingNet exemplar, a reference implementation for research on self-adaptation in the domain of IoT. DingNet offers a simulator that maps directly to a physical IoT system that is deployed in the area of Leuven, Belgium. DingNet models a set of geographically distributed gateways, which are connected to a user application that is deployed at a front-end server. The gateways can interact over a LoRaWAN network with local stationary and mobile motes that can be equipped with sensors and actuators. The exemplar comes with a set of scenarios for comparing the effectiveness of different self-adaptive solutions. We illustrate how the exemplar is used for a typical adaptation problem of smart city IoT application, where mobile motes dynamically have to adapt their communication settings to ensure reliable and energy efficient communication.
DingNet Paper