Scheduling Architectures for Scientific Workflows in the Cloud (Position Paper)

Scientific workflows describe a sequence of tasks that together form a scientific experiment. When workflows are computation or data intensive, distributed systems are used. Especially, cloud computing has gained a lot of attention due to its flexible and scalable nature. However, most approaches set up a preconfigured computation clusters or schedule tasks to existing resources. In this paper, we propose the utilization of cloud runtime models and couple them with scientific workflows to create the required architecture of a workflow task at runtime. Hereby, we schedule the architecture state required by a workflow task in order to reduce the overall amount of data transfer and resources needed. Thus, we present an approach that does not schedule tasks to be executed on resources, but schedule architectures to be deployed at runtime for the execution of workflows.