In our submission, we propose WiLD, an ontology to specify workflows models and instances in Linked Data, and present operational semantics for WiLD. Here, we show the five workflow models using which we evaluate our approach.
Interoperability between the plethora of devices on the market is a big obstacle for building applications for the Internet of Things. While the uniform interface of Linked Data (as put forward by the W3C’s Web of Things working group) paves the way towards combining such decentralised devices into integrated applications, traditional solutions for specifying what applications do, do not work seamlessly with Linked Data. We therefore tackle the problem of the specification, execution, and monitoring of applications in the environment of Linked Data. Understanding that workflows fit integration environments, we thus present a novel approach that combines workflows to specify applications with REST (for interaction integration) and semantic querying and reasoning (for data integration). We contribute to the state of the art by tailoring (1) an ontology for describing workflow models and instances, (2) operational semantics for the ontology, to the requirements of execution and monitoring in the environment of Linked Data. Moreover, we present a synthetic benchmark, where our approach shows linear behaviour in the size of the workload. Additionally, we showcase how we used our approach in practice in a Cyber-Physical System integration scenario to monitor pilots executing workflows in virtual aircraft cockpits.
We present the five workflow models for evaluating workflow engines determined by Ferme et al. in an analysis of workflow models found on the web, the literature, and industry (Ferme et al. (2017)). For our evaluation, we executed workflow instances of the workflow models described using the WiLD ontology. Here, we give those descriptions in RDF together with the workflow model depictions. To hint at applicability of the workflows in the Smart Building scenario of the benchmark, we give an interpretation for the workflow in scenario. Note that the WiLD ontology uses a tree-shaped workflow representation, to which we converted the flow-based workflows (as depicted).
