Scientific Workflow Systems

A challenge in many scientific experiments is managing the complexity of the experiment and the flows of data that are generated in all its phases. Most problem solving environments and virtual laboratories are built on top of scientific workflow systems, some of which exploit high-level scripting languages to define an experiment plan notation. Our objective is to research methods which enable building environments for collaborative planning, development and execution of e-Science applications, in particular:

• understanding mechanisms governing complex infrastructures on different levels to be able to efficiently support the whole life cycle of scientific experiments,

• investigation of new methods for managing the development and execution of complex scientific experiments and enhancing their life cycle by enabling re-usability and workflow interoperability,

• methods for the semantic integration of applications based on domain ontologies,

• approaches to collecting provenance data that will enable experiment reproducibility,

• development of a meta-level workflow scheduling to provide QoS and robustness for workflow execution.

Workflow systems, semantics, provenance, QoS.