Abstract: High energy consumption of the data center is a key problem to be solved in development of cloud computing. A cluster energy consumption optimization model and a Ceph cluster data-place policy based on the model are proposed. The combination of sequence and random storage methods is adopted while considering the cluster availability and fault tolerance, to reduce the data center energy consumption under the condition of satisfying the requirement of user’s Service-Level Agreement（SLA） and ensuring the clustering performance. The experimental results show that compared with the original Ceph cluster, the proposed data placement method makes the data center energy consumption reduced by 14.3%.

Key words: Ceph cluster, energy consumption model, data-place policy, energy consumption management

摘要： 数据中心能耗高是云计算发展过程中一个亟待解决的关键问题。建立了一个集群能耗优化模型，并在此基础上提出了一种基于Ceph集群的数据副本放置策略。该策略在考虑集群可用性和容错性的同时，采用了顺序存储和随机存储相结合的存储方法，使得Ceph集群在满足用户SLA需求和保证集群性能的前提下，达到节能的目的。实验结果表明，与原始Ceph集群相比，该数据副本放置策略在保证集群服务质量的同时，使得数据中心耗电量降低了14.3%。

关键词: Ceph集群, 能耗模型, 数据放置策略, 能耗管理