User-interface improvement and data provenance support for cloud-based scientific workflows

Abstract

Scientific workflows, which are organised as sets of tasks and data with dependencies, are used by scientists to model processes and perform simulated experiments. Scientific workflow management systems (SWMSs) enable scientists to design and execute scientific workflows with little concern for computational resource management. In addition, SWMSs are also responsible for tracking provenance data, which is the relationship between tasks and data, to enhance reproducibility of experiments and facilitate comparison of experiment outputs. Argo Workflows is one such SWMS based on Kubernetes, which can be easily deployed on any cloud computing platform for effortless resource management. However, the official user interface (UI) service of Argo Workflows is problematic, in terms of configurability of input data of workflows, comprehensiveness of data provenance support, and complexity of UIs. This project aims to address the issues in the official UI of Argo Workflows by implementing a new web portal application using React.js and a middleware application using Django. The functional requirements of the two applications are specified, and the detailed implementation is elaborated, which meets all the requirements and addresses all defects in the official UI. Usage examples are also provided to further demonstrate the implemented functionalities of the two applications.