项目名称: 基于混合存储介质的云存储环境的元数据管理和数据布局研究
项目编号: No.61303073
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 自动化技术、计算机技术
项目作者: 陈涛
作者单位: 中国人民解放军军事医学科学院
项目金额: 23万元
中文摘要: 基于云存储对固态存储介质的需求以及固态存储介质和磁盘存储介质各自的优 势,构建高效能、大容量、多介质、多层次的基于混合存储介质的云存储是存储系统的主要 发展趋势之一。现有的元数据管理和数据布局是面向基于磁盘的传统存储系统,没有考虑固 态存储的特性,在基于混合存储介质的云存储环境中无法充分发挥异质存储资源的优势。 本项目面向混合存储介质的云存储环境,研究和突破不同存储介质和层次结构的元数据 管理和数据布局关键技术:研究面向混合存储介质的云存储环境的可扩展的体系结构,以此 结构为基础,探索有效的元数据冗余分配算法和动态、自适应的元数据负载均衡机制,突破 基于数据访问行为的动态、自适应的数据布局和面向混合存储介质特性的公平、自适应的副 本布局等技术,提高存储的I/O 性能,自适应存储规模的变化,提高数据存储的可用性和可 靠性,为基于混合存储介质的云存储环境的数据管理提供理论方法和技术支撑。
中文关键词: 云存储;固态存储;混合存储;元数据管理;数据布局
英文摘要: According to the demand of solid-state storage media for cloud storage, and the respective advantages of solid-state storage media and disk storage media, building high-performance, large capacity, multi-media, multi-level cloud storage with hybrid storage media is one of the major trends of the storage system. Existing metadata management and data placement is oriented to the characteristics of traditional disk-based storage systems, without considering the characteristics of solid-state storage. Thus, the existing approaches can not give full play to the advantages of heterogeneous storage resources in the cloud storage environment with hybrid storage media. Based on cloud storage environment with hybrid storage media, this project focuses on the key technologies of metadata management and data placement in different storage media: scalable architecture of cloud storage environment with hybrid storage media, effective and redundant metadata allocation algorithm, dynamic and adaptive metadata load balancing mechanism, dynamic and adaptive data placement with behavior of data access, fair and adaptive replica placement with the characteristics of hybrid storage media. This goal of the project is to improve storage I/O performance, adapt to the changes of storage devices, and improve the availability and reliabil
英文关键词: cloud storage;solid-state storage;hybrid storage;metadata management;data placement