DynaSched: a dynamic Web service scheduling and deployment framework for data-intensive Grid workflows

Abstract Grid computing boosts productivity by maximizing resource utilization and simplifying access to resources which are shared among virtual organizations. Recently, the Grid and Web Service communities have established a set of common interests and requirements. The latest version of the Globus Toolkit implements the Web Service Resource Framework (WSRF) specifications which have been formulated to cover these interests. We leverage the Globus Toolkit to address some limitations in supporting the dynamic nature of large-scale Grid and data-intensive workflow executions. Dynamic Web Service deployment fits well into the dynamic nature of the Grid and opens new ways of managing workflow executions on the Grid. In this article, we present the design and evaluation of a dynamic Web Service scheduling and deployment framework (DynaSched) that supports the workflow management of dynamic services. Dynamic Web Service deployment on the Grid allows jobs to be executed on the same site as where the input data is located. The empirical studies show that the designed framework decreases data-intensive workflow execution time by minimizing communication costs. We argue that the framework ensures more flexible, fault-tolerant workflows. The system is based on Open Grid Services Architecture specifications and is WSRF-compliant.