Multiple Workflow-based Energy-efficient Task Scheduling Research In Cloud Computing Environment

Workflow management system in cloud environment, is suitable for the field of cooperation in need of mass data storage, transmission and efficient computing performance across the organizations and businesses and is gradually playing an irreplaceable role in physics, biology, geography, meteorology, astronomy, mathematics, and many other fields.With energy saving management becoming an important subject in cloud green computing, energy-saving scheduling ideology of workflow caused wide public concern because of its low cost and reliability. At present, the research based on the workflow dependent task scheduling in cloud environment has a great challenge to conserve energy, it needs to balance the conflicting relationship in task priority binding, load balancing, huge amounts of data transmission, total completion time of the tasks and other performance indicators. In the case of a large number of concurrent workflow instances, when the workflow tasks were performed on different resources, and need the data of different location, especially the data having dependent constraints between tasks, how to effectively schedule these tasks and resources and optimize the task data transmission has become the core issue which must be addressed. Traditional energy saving scheduling strategies focus on how to reduce the task completion time, ignoring the resource utilization, or focus on the research how to solve the contradiction between reducing task completion time and improving the resource utilization, ignoring the requirements of energy saving, which makes the workflow scheduling results severely affected and user satisfaction reduced.According to the research and analysis on existing scientific workflow task scheduling strategy in the cloud environment, a task scheduling algorithm based on multiple workflows(MWFTS) was proposed to dispose the easily appearing idle resources phenomenon of SHEFTEX in dealing with the multiple workflow, and for this scheduling algorithm, this paper proposes an energy saving strategy(Multiple Workflow Task Energy- Efficient Scheduling, MWFTEES). Firstly, all the tasks in these workflows were integrated together to be evaluated,and then were sorted to a task execution sequence. Then for each task,in the sequence,a resource with the earliest execution completion time were selected to execute the task after iterating through all the resources. Before the current task execution, it first checked if the allocated resource had free time. If it did have, the processor frequency could be adjusted for the previous task allocated on the resource, thus saving energy.The experimental simulation shows that MWFTS has less completion time and higher resource utilization than SHEFTEX; MWFTEES strategy has the same completion time with MWFTS, but higher resource utilization and less energy consumption than MWFTS.