Energy-Aware Workflow Scheduling in Grid Under QoS Constraints

In this study, we have considered the problem of scheduling precedence-constraint parallel applications (workflows) in heterogeneous grid-computing environment. Recently many heuristics have been devoted to grid scheduling typically restricted to optimizing the execution time (makespan) only without paying much concentration on energy consumption. Reducing energy consumption can bring various advantages like reducing operating costs, environmental perspective and increase in system reliability. This paper aims to develop energy-aware task scheduling algorithm in grid based on the dynamic voltage and frequency scaling (DVFS) technique. The user negotiates with the service provider on their quality of service (QoS) requirements along with green computing specifications to reach the service level agreement. With the use of DVFS, the algorithm minimizes the energy consumption of task execution while satisfying the QoS constraints (deadline). The proposed static scheduling algorithm works in three phases: deadline distribution, tasks ordering and then assigning the best services to tasks along with selecting the appropriate voltage levels while meeting its sub-deadline. The simulation results using randomly generated task graphs and task graphs corresponding to real-world problems exhibit that the proposed algorithm achieves energy efficiency and reduces energy consumption up to 68 % with the increase in 30 % of the execution time.