Workflow Scheduling on Federated Clouds

Federated Clouds, or the orchestration of multiple Cloud services for fulfilling applications' requirements, is receiving increasing attention. Despite their many advantages, federated Clouds also present some downsides since different services may reside in different geographically located areas. This paper focuses on evaluating the advantages and disadvantages, from the point of view of performance and financial costs, of using a federation of Clouds for executing scientific workflows. It evaluates a wide range of different workflow types with different requirements in terms of computation and communication (produced and consumed data), and discusses which kind of workflow applications can benefit from a Cloud federation and how.