Due to their abundant resources that can be elastically provisioned with pay-as-you-go pricing, clouds have emerged as a promising cost-efficient platform to execute large scale scientific applications. Such applications consist of number of processes/tasks forming workflow. These tasks are connected by direct edges that show the data dependency between the tasks. Tasks perform their computation on the original data submitted by the user, or on data passed by its predecessor task. This work, classify and discuss proposals that investigate the problem of scheduling scientific workflows in clouds.
Alrawashdeh et al. (Tue,) studied this question.