首页 | 本学科首页   官方微博 | 高级检索  
     

LD_B~+树:闪存系统上一种DBMS索引结构
引用本文:李国徽,赵培,高胜. LD_B~+树:闪存系统上一种DBMS索引结构[J]. 小型微型计算机系统, 2012, 33(2): 193-198
作者姓名:李国徽  赵培  高胜
作者单位:华中科技大学计算机科学与技术学院,武汉,430074
基金项目:国家自然科学基金项目(60873030)资助;国家“八六三”高技术研究发展计划项目(2007AA01Z309)资助;国家“十一五”国防预研基金资助
摘    要:以往的研究大多针对文件系统,而DBMS存在更多细粒度的更新.本文综合考虑闪存自身的特点、设备种类繁多及不同闪存设备读写特性差别大等,提出了一种基于闪存的DBMS索引结构:LD_B+树.LD_B+树根据工作负载的读写特性动态地调节索引模式使之能够适应于不同种类的闪存设备.LD_B+树采用日志结构组织结点,通过结点转换表和日志缓冲区维护索引结构.模拟实验结果表明,不同闪存设备及工作负载下,LD_B+索引结构比B+树和日志型B+树(BFTL)具有6%-63%的性能提高.

关 键 词:LD_B+树  日志  结点转换表  Merge操作  在线算法  恢复

LD_B~+ Tree: DBMS Index Structure over Flash Memory Storage Systems
LI Guo-hui , ZHAO Pei , GAO Sheng. LD_B~+ Tree: DBMS Index Structure over Flash Memory Storage Systems[J]. Mini-micro Systems, 2012, 33(2): 193-198
Authors:LI Guo-hui    ZHAO Pei    GAO Sheng
Affiliation:(College of Computer Science and Technology,Huazhong University of Science and Technology,Wuhan 430074,China)
Abstract:Flash memory has been widely used in the storage system.Most researches mainly focuses on the file system,However,DBMS has more fine-granularity updates.Considering a wide range of flash memory devices have far different access characteristics,a new DBMS index structure,referred to as LD_B+ tree,is proposed in the paper,which redefines the basic indexing operation such as retrieval,insertion,deletion,and so on.LD_B+ tree mainly uses node translation table and logging buffer to maintain the indexing structure.A set of simulation experiments was conducted and the results showed that the LD_B+ tree outperformed the existing index structure 6%-63% under different workloads and flash packages.
Keywords:LD_B+ tree  logging unit  node translate table  merge  online algorithm  recovery
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号