Distributed Scientific Workflow Management for Data-Intensive Applications

Grids and service-oriented technologies are emerging as dominant approaches for distributed systems. Rising complexity of grid applications places new requirements and increases needs to improve the design and reusability of grid workflow systems. In particular, scientific data-intensive workflows require new approaches for integration and coordination of distributed resources. Traditional centralized approaches for workflow execution can be quite inefficient for such workflows. In order to clarify these issues, the paper surveys existing workflow management systems evaluating them practically on some use cases for management of natural resources from the multidisciplinary research project GLOWA Volta. Subsequently, the most important obstacles are identified and a new approach facilitating semantic oriented composition, reuse and distributed execution of workflows is proposed.