Abstract: According to the users’ requirements of QoStime and execution costs， the scheduling strategies for workflow were modified by combining the cloud computing and workflow. For the problem that workflow efficiency of instanceintensive cloud workflows was not high in a cloud environment， the cloud workflow scheduling model and the modified scheduling algorithm for cloud workflow were proposed based on QoS. DAG was used in the algorithm， and resource strategy was modified to meet the latest execution time under the premise of minimizing the entire workflow execution costs. The experimental results showed that the efficiency effectively improved and the cost of using resource was reduced with the proposed algorithm.

Key words: cloud workflow, QoS, resource optimization, DAG(directed acyclic graph), time deadline