Date Added: Apr 2010
The development of cloud computing has generated significant interest in the scientific computing community. In this paper, the authors consider the impact of cloud computing on scientific workflow applications. They examine the benefits and drawbacks of cloud computing for workflows, and argue that the primary benefit of cloud computing is not the economic model it promotes, but rather the technologies it employs and how they enable new features for workflow applications. They describe how clouds can be configured to execute workflow tasks, and present a case study that examines the performance and cost of three typical workflow applications on Amazon EC2. Finally, they identify several areas in which existing clouds can be improved and discuss the future of workflows in the cloud.