Power Consumption Optimization Strategy of Cloud Workflow Scheduling Based on SLA

摘要

Cloud computing, as a new model of service provision in distributed computing environment, faces the great challenge of energy consumption because of its large demand for computing resources. Choosing improper scheduling method to execute cloud workflow tends to result in the waste of power consumption. In order to lower the higher power consumption for cloud workflow executing, we propose a power consumption optimization algorithm for cloud workflow scheduling based on SLA (Service Level Agreement), which can reduce power consumption while meeting the performance-based constraints of time and cost. The algorithm first searches for all feasible scheduling solutions of cloud workflow application with critical path, then the optimal scheduling solution can be found out through calculating total power consumption for each feasible scheduling solution. The experimental results show that compared with traditional workflow scheduling algorithms based on QoS, the optimization algorithm proposed in this paper not only meets the constraints of time and cost defined in SLA, but also reduces the average power consumption by around 10%.