鉴于云计算存在庞大的节点规模和数据规模，云计算系统在可扩展性、高可用性等方面面临着诸多问题。针对云计算环境下的分布数据存储，构建了一种基于递归的正N边形网络模型，在数据中心网络结构的基础上给出了数据管理模型，并设计了以提高数据可用性、负载平衡为目的的副本分布策略和副本选择策略。在保证数据可用性的前提下，提出一种基于覆盖集的数据迁移算法，利用节点选择策略尽量减少迁移成本，进行数据迁移让更多的机器休眠，降低能耗。通过对实验数据进行比较分析，验证了提出的网络拓扑结构、数据管理模型、数据迁移技术的正确性和有效性。%The cloud computing system is faced with many problems in the aspects of scalability and high availability due to the enormous node scale and data scale of cloud computing. For the distributed data storage in cloud computing environment, an regular N⁃equilateral polygon(RNEP)network model based on recursion was constructed. The data management model is given on the basis of data center network structure. The transcript distribution strategy and selection strategy were designed to improve the data availability and load balancing ability. A data migration algorithm based on covering set is proposed on the premise of ensuring the data availability,in which the node selection strategy is used to minimize the migration cost and conduct data mi⁃gration to rest more machines and reduce energy consumption. The experimental data is compared and analyzed,which proves that the proposed network topology,data management model and data migration technology are valid and effective.
展开▼
