项目名称: 高可靠易扩展的固态盘阵列关键技术研究
项目编号: No.61272055
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 张广艳
作者单位: 清华大学
项目金额: 86万元
中文摘要: 青年项目取得一些重要进展,首次发现并证明了循环RAID扩展过程中存在可重排序窗口特性,提出了循环RAID-5快速扩展方法ALV;打破数据循环分布的约束,在保证数据各盘均匀分布的同时最小化了数据迁移量,提出了RAID-0快速扩展方法FastScale;提出了RAID-5快速扩展方法MiPiL。主要学术成绩包括:申请发明专利1项,发表学术论文5篇,其中1篇发表在存储领域最好的国际会议FAST’11上,1篇发表在系统领域国际顶尖期刊IEEE TC上,SCI检索2篇,EI检索4篇。这些研究使得申请者对磁盘阵列深入理解,对Flash介质特征相当熟悉。在此基础上,申请者提出高可靠易扩展的固态盘阵列关键技术研究,研究内容包括:校验更新对固态盘寿命损耗的消除技术;固态盘阵列数据优化布局技术;固态盘阵列快速在线重构技术;固态盘阵列快速在线扩展技术。最终目标是实现高性能、高可靠、易扩展的固态盘阵列。
中文关键词: 固态盘;冗余阵列;可靠性;扩展性;数据布局
英文摘要: Under the support of the youth project, we make some important achievements as follows. We discover and prove the characteristics of reordering window during scaling a round-robin RAID; propose an efficient scaling approach for round-robin RAID-5 called ALV; break the restriction of round-robin data distribution to make sure an even data distribution among different disks with minimal data migration; propose an efficient scaling approach for RAID-0 called FastScale; propose an efficient scaling approach for RAID-5 called MiPiL. The academic achievements include one patent and five papers. Among these papers, one appears in a top-level conference in the storage area (FAST’11), another appears in a top-level journal (IEEE Transactions on Computers). Two of them are indexed by SCI and four of them are indexed by EI. With these research experiences, the applicant understands deeply disk arrays and is familiar to the characteristics of flash media. On this basis, the applicant proposes to research key technologies in reliable and scalable solid-state disk arrays. The research contents include: eliminating the negative impact of maintaining parity on the solid-state disk lifespan, optimizing data layouts in solid-state disk arrays, reconstructing solid-state disk arrays quickly, and scaling solid-state disk arrays qui
英文关键词: solid-state drive;redundant array;reliability;scalability;data layout