Combining local and grid resources in scientific workflows (for Bioinformatics)
Ann-Charlotte Berglund, Erik Elmroth, Francisco Hernández, Björn Sandman, Johan Tordsson · 2009
Abstract. We examine some issues that arise when using both local and Grid resources in scientific workflows. Our previous work addresses and illustrates the benefits of a light-weight and generic workflow engine that manages and optimizes Grid resource usage. Extending on this effort, we here illustrate how a client tool for bioinformatics applications employs the engine to interface with Grid resources. We also explore how to define data flows that transparently integrates local and Grid subworkflows. In addition, the benefits of parameter sweep workflows are examined and a means for describing this type of workflows in an abstract and concise manner is introduced. Finally, the above mechanisms are employed to perform an orthology detection analysis.