Freezing Frozen Pages with Multi-Stream SSDs

Hyunwoo Park, Soyee Choi, Mijin An, Sang Won Lee
{"title":"Freezing Frozen Pages with Multi-Stream SSDs","authors":"Hyunwoo Park, Soyee Choi, Mijin An, Sang Won Lee","doi":"10.1145/3329785.3329935","DOIUrl":null,"url":null,"abstract":"Flash memory SSDs have been replacing hard disks as the main storage thanks to their high IOPS/$ and energy e ciency [6]. Internally within SSD, Flash Translation Layer (FTL) has control over data mapping and garbage collection (GC). Because of its architectural feature, log-structured approach, the costly GC operation is unavoidable to secure new blocks for newly written pages. During every GC, valid pages from the victim block have to be relocated causing additional writes. It is well known that these additional writes (i.e., write ampli cation) negatively a ect the performance and lifespan of the ash memory SSDs. Hence, it is imperative to reduce write ampli cation factor (WAF) in modern ash memory SSDs. Though numerous FTLs have been proposed with this goal in mind, the most commercial ash storage device is known to take the page-mapping FTL approach [7]. Unfortunately, it is well known that page-mapping FTL results in highWAF (e.g., 5), particularly against random write patterns which are common in OLTP workloads. Recently, one interesting interface for ash memory SSDs, called Multi-Streamed SSD (MS-SSD in short), has been proposed and standardized to lower the GC overhead with the help of explicit hint on data lifetime from applications [4, 8]. The main purpose of MS-SSD is to allow to place data pages with di erent lifetime into di erent ash regions so as to reduce the SSD-internal WAF from GC. Meanwhile, the e ectiveness of MS-SSD is highly dependent on the accuracy of the classi cation of the lifetime of each data pages [3]. Therefore, it is critical for any MS-SSD applications to precisely classify logical data pages according to their","PeriodicalId":376796,"journal":{"name":"Proceedings of the 15th International Workshop on Data Management on New Hardware","volume":"2 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"5","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings of the 15th International Workshop on Data Management on New Hardware","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/3329785.3329935","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 5

Abstract

Flash memory SSDs have been replacing hard disks as the main storage thanks to their high IOPS/$ and energy e ciency [6]. Internally within SSD, Flash Translation Layer (FTL) has control over data mapping and garbage collection (GC). Because of its architectural feature, log-structured approach, the costly GC operation is unavoidable to secure new blocks for newly written pages. During every GC, valid pages from the victim block have to be relocated causing additional writes. It is well known that these additional writes (i.e., write ampli cation) negatively a ect the performance and lifespan of the ash memory SSDs. Hence, it is imperative to reduce write ampli cation factor (WAF) in modern ash memory SSDs. Though numerous FTLs have been proposed with this goal in mind, the most commercial ash storage device is known to take the page-mapping FTL approach [7]. Unfortunately, it is well known that page-mapping FTL results in highWAF (e.g., 5), particularly against random write patterns which are common in OLTP workloads. Recently, one interesting interface for ash memory SSDs, called Multi-Streamed SSD (MS-SSD in short), has been proposed and standardized to lower the GC overhead with the help of explicit hint on data lifetime from applications [4, 8]. The main purpose of MS-SSD is to allow to place data pages with di erent lifetime into di erent ash regions so as to reduce the SSD-internal WAF from GC. Meanwhile, the e ectiveness of MS-SSD is highly dependent on the accuracy of the classi cation of the lifetime of each data pages [3]. Therefore, it is critical for any MS-SSD applications to precisely classify logical data pages according to their
使用多流固态硬盘冻结冻结页面
由于其高IOPS/$和能源效率,闪存ssd已经取代硬盘成为主要存储[6]。在SSD内部,Flash转换层(FTL)控制数据映射和垃圾收集(GC)。由于其体系结构特性,即日志结构化方法,为确保新写入页面的新块的安全,代价高昂的GC操作是不可避免的。在每次GC期间,必须重新定位来自受害块的有效页面,从而导致额外的写操作。众所周知,这些额外的写入(即写入放大)会对灰存储ssd的性能和寿命产生负面影响。因此,降低现代灰存储固态硬盘的写入放大系数(WAF)势在必行。尽管已经提出了许多FTL以实现这一目标,但已知最商业化的灰存储设备采用页面映射FTL方法[7]。不幸的是,众所周知,页面映射FTL会导致高waf(例如,5),特别是针对OLTP工作负载中常见的随机写模式。最近,人们提出了一种有趣的灰内存SSD接口,称为多流SSD(简称MS-SSD),并对其进行了标准化,通过明确提示应用程序的数据生命周期来降低GC开销[4,8]。MS-SSD的主要目的是允许将具有不同生命周期的数据页放置到不同的灰区,从而减少来自GC的ssd内部WAF。同时,MS-SSD的有效性高度依赖于每个数据页生命周期分类的准确性[3]。因此,对于任何MS-SSD应用程序来说,根据它们的逻辑数据页精确地分类是至关重要的
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信