Semantic-Based Scientific Workflow Composition

Scientific workflow composition is still a challenging task for domain scientists. One of the main reasons is the lack of domain knowledge support in workflow composition tools or languages. Existing knowledge-based scientific workflow languages and composition tools require sophisticated expertise from domain scientists in order to automate the process of managing workflows and their components (activities). This chapter presents a novel semantic-based approach which features separation of concerns between data semantics and data representation, and between Activity Function (AF) (semantic description of workflow activities) and Activity Type (AT) (syntactic description of workflow activities). The separation, implemented as part of AWDL, simplifies scientific workflow composition by enabling users to compose scientific workflows at the level of data semantics and Activity Function (AF), as well as providing for automatic data conversion (Chap. 7). The separation also provides a foundation to build effective algorithms for automatic scientific workflow composition (Chap. 8). We implemented our approach as part of ASKALON and demonstrated its effectiveness by applying it to a real-world meteorology workflow application. Our approach can also be adapted to other scientific domains by developing the corresponding ontologies for those domains.