Reducing Energy Consumption With Cost Budget Using Available Budget Preassignment in Heterogeneous Cloud Computing Systems.

Energy consumption and cost are the important factors in executing applications in grid or cloud computing systems, because they directly affect resource consumption and economic benefits. This paper solves the problem of reducing energy consumption of a cost budgeted directed acyclic graph (DAG) application in heterogeneous computing systems. The state-of-the-art work has studied the cost budgeted scheduling for a DAG application in the heterogeneous computing systems by proposing the budget level based preassignment method; however, this paper is merely to reduce the schedule length without involving energy consumption, and its preassignment method is still pessimistic although it improves the existing method. In this paper, we first propose an available budget preassignment method to further improve it. We then introduce the available budget preassignment method to reduce the energy consumption. We propose minimizing energy consumption using the available budget preassignment (MECABP) algorithm based on the two steps. Experiments on three types of DAG applications with different parallelism degrees confirm the effectiveness of the proposed MECABP algorithm compared with existing algorithms.