中文摘要：

日益增长的数据中心冷却系统能耗会提升数据中心运行成本并带来巨大的碳排放,在数据中心轻载时进行负载聚合是降低系统能耗的有效途径.传统负载聚合只关注计算功耗的降低,而负载聚合会带来数据中心热点的产生和冷却功耗的增加.通过热力学仿真得到温度功耗模型;以一个实际云计算平台内的工作日志作为负载进行数据中心仿真,结果表明考虑节点热传递特性的聚合算法比起考虑迁移开销的聚合算法更能有较降低峰值温度和冷却功耗.

英文摘要：

The ever-growing energy cost especially cooling cost of data centers has been emphasized much for carbon emissions reduction.Workload consolidation could be applied to data centers for energy reduction.Traditional methods only focused on computing energy reduction,while this paper puts emphasis on hot spot brought by work consolidation and further cooling cost increasing brought by work consolidation.The power temperature model of data center could be constructed based on heat recirculation matrix,which could be got through thermal dynamic simulation.A data center with the working log got from a real cloud computing platform was simulated.The experiment results show that the algorithms which considering heat recirculation among nodes have less peak inlet temperature,cooling cost and harm on performance than algorithm considering only migration cost in different conditions.