PROJECT TITLE :

Energy Utilization Task Scheduling for MapReduce in Heterogeneous Clusters

ABSTRACT:

At this point in time, the cost of energy is the most important consideration in cloud computing. Therefore, it is of the utmost importance to implement methods of task scheduling that take into account energy efficiency. For the purpose of cutting down on energy expenses in heterogeneous clusters, a task scheduling framework that takes into account deadlines, data locality, and resource utilization has been proposed. The construction of the task list, the scheduling of the task, and the updating of the slot list make up the framework. A new job sequence is proposed to construct a reasonable task list, taking into consideration the limitations imposed by the deadlines, the number of job slots that have been allotted, and the possible processing times of the jobs. In the produced task scheduling, tasks are scheduled to promising slots from their rack-local servers, cluster-local servers, and remote servers. This significantly improves the degree to which data is located locally. After the tasks and slots have been distributed, it is suggested that an update of available slots in clusters be performed. This would not only help to locate vacant slots, but it would also improve the server's resource utilization by making use of fuzzy logic to adjust the number of available slots in accordance with the amount of bandwidth, memory, and CPU that is being used. The proposed heuristic uses significantly less energy than the modified versions of existing algorithms that have a variable total number of slots, according to the findings of the experiments that were conducted.