ABSTRACT
The number of applications needing big data is on the rise nowadays, where the big data processing tasks are sent as workflows to cloud computing systems. Considering the recent advances in the Internet technology, cloud computing have become the most popular computing technology. The scheduling approach in cloud computing environments has always been a topic of interest to many researchers. This paper proposes a new scheduling algorithm for data-intensive workflows based on data dependencies in computational clouds. The proposed algorithm tries to minimize the makespan by considering the details of the workflow structure and virtual machines. The concepts and details defined and considered in this study has received less emphasis in previous works. According to the results, the proposed algorithm reduced the duration of communication between tasks and runtimes by taking into account the features of data-intensive workflows and proper task assignment. Consequently, it reduced the total makespan in comparison with previous algorithms.
Key words: Bottleneck task, Data-intensive applications, Resource allocation, Sensitive task, Workflow scheduling
|
