A Grid simulation framework to study advance scheduling strategies for complex workflow applications

Workflow scheduling in Grids becomes an important area as it allows users to process large scale problems in an atomic way. However, validating the performance of workflow scheduling strategies in real production environment cannot be feasibly carried out. The complexity of production systems, dynamicity of Grid execution environments, and the difficulty to reproduce experiments, make workflow scheduling production systems a complex research environment. Instead, this work is based on a trace driven simulator. This work presents workflow scheduling support as an extension to the Teikoku Grid Scheduling Framework (tGSF). tGSF was developed as a response for a standard compliant and trace based Grid scheduling simulation environment. Workflow scheduling is provided via a second layer of Grid scheduler, extensible to new workflow and parallel scheduling strategies. This work also includes a usage case scenario, which illustrates how this extension can be used for quantitative experimental study.