{"title":"RS-Wrapper: random write optimization for solid state drive","authors":"Da Zhou, Xiaofeng Meng","doi":"10.1145/1645953.1646144","DOIUrl":null,"url":null,"abstract":"Solid State Drive (SSD), emerging as new data storage media with high random read speed, has been widely used in laptops, desktops, and data servers to replace hard disk during the past few years. However, poor random write performance becomes the bottle neck in practice. In this paper, we propose to insert unmodified data into random write sequence in order to convert random writes into sequential writes, and thus data sequence can be flushed at the speed of sequential write. Further, we propose a clustering strategy to improve the performance by reducing quantity of unmodified data to read. After exploring the intrinsic parallelism of SSD, we also propose to flush write sequences with the help of the simultaneous program between planes and parallel program between devices for the first time. Comprehensive experiments show that our method outperform the existing random-write solution up to one order of magnitude improvement.","PeriodicalId":286251,"journal":{"name":"Proceedings of the 18th ACM conference on Information and knowledge management","volume":"57 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2009-11-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"6","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 18th ACM conference on Information and knowledge management","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/1645953.1646144","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 6
Abstract
Solid State Drive (SSD), emerging as new data storage media with high random read speed, has been widely used in laptops, desktops, and data servers to replace hard disk during the past few years. However, poor random write performance becomes the bottle neck in practice. In this paper, we propose to insert unmodified data into random write sequence in order to convert random writes into sequential writes, and thus data sequence can be flushed at the speed of sequential write. Further, we propose a clustering strategy to improve the performance by reducing quantity of unmodified data to read. After exploring the intrinsic parallelism of SSD, we also propose to flush write sequences with the help of the simultaneous program between planes and parallel program between devices for the first time. Comprehensive experiments show that our method outperform the existing random-write solution up to one order of magnitude improvement.
近年来,固态硬盘(Solid State Drive, SSD)作为一种具有高随机读取速度的新型数据存储介质,被广泛应用于笔记本电脑、台式机和数据服务器中,以取代硬盘。但在实际应用中,随机写性能差成为瓶颈。在本文中,我们提出将未修改的数据插入到随机写序列中,将随机写转换为顺序写,从而可以以顺序写的速度刷新数据序列。此外,我们提出了一种聚类策略,通过减少未修改数据的读取量来提高性能。在探索了SSD的内在并行性之后,我们还首次提出了利用平面间并行编程和设备间并行编程来刷新写入序列。综合实验表明,我们的方法比现有的随机写入解决方案提高了一个数量级。