For the data-intensive scientific workflows, large mount of data need to be stored and often been distributed into several data centers. So the data movements between different data centers become inevitable when the application executed. The more time the data movements, the less efficiency of the data-intensive application. This paper proposed a data placement method based on Bayesian network for data-intensive scientific workflows, which could effectively reduce the data movement time between different data centers. The validity of this method is discussed and illustrated by experiment.
展开▼
