›› 2013,Vol. 28 ›› Issue (1): 28-41.doi: 10.1007/s11390-013-1310-z

所属专题： Computer Architecture and Systems

• Special Section on Selected Paper from NPC 2011 • 上一篇    下一篇

一种提高基于SSD的RAID系统的寿命和性能的盘间损耗均衡方法

Yi-Mo Du (杜溢墨), Nong Xiao* (肖侬), Member, IEEE, Fang Liu (刘芳), Member, CCF and Zhi-Guang Chen (陈志广)   

  • 收稿日期:2011-12-31 修回日期:2012-03-30 出版日期:2013-01-05 发布日期:2013-01-05
  • 基金资助:

    Supported by the National High Technology Research and Development 863 Program of China under Grant No. 2013AA013201, the National Natural Science Foundation of China under Grant Nos. 61025009, 61232003, 61120106005, 61170288.

CSWL: Cross-SSD Wear-Leveling Method in SSD-Based RAID Systems for System Endurance and Performance

Yi-Mo Du (杜溢墨), Nong Xiao* (肖侬), Member, IEEE, Fang Liu (刘芳), Member, CCF and Zhi-Guang Chen (陈志广)   

  1. State Key Laboratory of High Performance Computing, National University of Defense Technology, Changsha 410073, China
  • Received:2011-12-31 Revised:2012-03-30 Online:2013-01-05 Published:2013-01-05
  • Supported by:

    Supported by the National High Technology Research and Development 863 Program of China under Grant No. 2013AA013201, the National Natural Science Foundation of China under Grant Nos. 61025009, 61232003, 61120106005, 61170288.

PDF (PC)

3587

随着技术的不断进步,闪存(Flash memory)以其低价格、高性能、非易失、低功耗和容量大等特点,已经从当初以数字家电、手机等为主的嵌入式领域应用逐步向服务器和高性能计算机以及大规模存储系统的应用转变.百度、google、以及facebook等公司已经在其系统中开始集成固态存储模块以增强服务处理性能.而基于闪存固态盘构架RAID阵列系统无疑是一种构建大规模固态存储系统的很好手段.虽然闪存应用日益普遍,但是还是存在几个制约其进一步发展的主要问题：1.非定点更新问题,即每次需要重写一页之前需要对该页所在的整块进行擦除;2.为了保证可靠性,对于块内页的写必须顺序依次进行;3.寿命问题,即所有块都有擦除次数限制.针对这些问题,学术界和产业界都给出了很多解决方案.本文主要针对第三个问题提出解决方案.当前市场上基于闪存的固态盘(Solid State Drive)基本都采用了两种方法来解决寿命问题,使得固态盘能够保证按其按所给的容量持续服务.一种是内部的损耗均衡策略避免固态盘中的某些块被快速磨损;另外一种是在提供基本容量以外还提供一些额外的空间来对一些因使用寿命或其它原因导致的坏块进行替换.由于闪存存在寿命问题,所以基于闪存的固态盘同样存在这个问题,但是目前还没有很好的基于固态盘阵列的盘间损耗均衡方案,这会导致系统中某些固态盘可能会先于其他盘达到使用寿命而被替换掉,替换后,数据需要重构,这是一个及其耗时的过程,对于系统的可靠性和可用性都会造成极大的影响.基于此,我们提出了一种基于固态盘的RAID系统的盘间损耗均衡方法CSWL.由于在随机访问模式下,基于纠删码的RAID系统中的检验数据比普通数据要被更频繁的更新,所以CSWL采用了基于损耗程度的动态检验信息分配方案：让损耗严重的盘分配较少的校验数据,让损耗较轻的盘分配较多的校验数据,这样可以达到延长整个系统寿命的目的.同时,由于校验信息的动态分布使得写负载重的盘由于校验数据的减少而负载减轻,而写操作又是制约固态盘性能的关键因素,这种对于写操作的负载均衡对性能的提高也有很大的帮助.

Abstract: Flash memory has limited erasure/program cycles. Hence, to meet their advertised capacity all the time, flash- based solid state drives (SSDs) must prolong their life span through a wear-leveling mechanism. As a very important part of flash translation layer (FTL), wear leveling is usually implemented in SSD controllers, which is called internal wear leveling. However, there is no wear leveling among SSDs in SSD-based redundant array of independent disks (RAIDs) systems, making some SSDs wear out faster than others. Once an SSD fails, reconstruction must be triggered immediately, but the cost of this process is so high that both system reliability and availability are affected seriously. We therefore propose cross-SSD wear leveling (CSWL) to enhance the endurance of entire SSD-based RAID systems. Under the workload of random access pattern, parity stripes suffer from much more updates because updating to a data stripe will cause the modification of other all related parity stripes. Based on this principle, we introduce an age-driven parity distribution scheme to guarantee wear leveling among flash SSDs and thereby prolong the endurance of RAID systems. Furthermore, age-driven parity distribution benefits performance by maintaining better load balance. With insignificant overhead, CSWL can significantly improve both the life span and performance of SSD-based RAID.

[1] Chen F, Luo T, Zhang X D. CAFTL: A content-aware flashtranslation layer enhancing the lifespan of flash memory basedsolid state drives. In Proc. the 9th FAST, Feb. 2011, pp.77-90.
[2] Patterson D, Gibson G, Katz R H. A case for redundant arraysof inexpensive disks (RAID). In Proc. the 1988 SIGMOD,June 1988, pp.109-116.
[3] Zhen W M, Zhang G Y. FastScale: Accelerate RAID scal-ing by minimizing data migration. In Proc. the 9th FAST,February 2011, pp.149-161.
[4] Balakrishnan M, Kadav A, Prabhakaran V, Malkhi D. Differ-ential RAID: Rethinking RAID for SSD reliability. In Proc.Eurosys, April 2010, pp.15-26.
[5] Grupp L M, Caulfield A M, Coburn J, Swason S, Yaakobi E,Seigel P H, Wolf J K. Characterizing flash memory: Anoma-lies, observations, and applications. In Proc. the 42nd MI-CRO, December 2009, pp.24-33.
[6] Thomasian A, Blaum M. Higher reliablity redundant disk ar-rays: Organizations, operation, and coding. ACM Transac-tion on Storage, 2009, 5(3): Article No.7.
[7] Du Y M, Xiao N, Liu F, Chen Z G. A customizable and mod-ular flash translation layer (FTL) design and implementation.Journal of Xi'an Jiaotong University, 2010, 44(8): 42-47. (InChinese)
[8] Gal E, Toledo S. Algorithms and data structures for flashmemories. ACM Computing Surveys, 2005, 37(2): 138-163.
[9] Park K, Lee D H, Woo Y et al. Reliability and performanceenhancement technique for SSD array storage system usingRAID mechanism. In Proc. the 9th ISCIT, Sept. 2009,pp.140-145.
[10] Lee D D, Seung S H. Algorithms for non-negative matrix fac-torization. Advances in Neural Information Processing Sys-tems, 2001, 13: 556-562.
[11] Geist R, Trivedi K. An analytic treatment of the reliabilityand performance of mirrored disk subsystems. In Proc. the23rd Inter. Symp. Fault-Tolerant Computing, June 1993,pp.442-450.
[12] Thomasian A. Shortcut method for reliability comparisons inRAID. Journal of Systems and Software, 2006, 79(11): 1599-1605.
[13] Mao B, Jiang H, Feng D,Wu S Z, Chen J X, Zeng L F, Tian L.HPDA: A hybrid parity-based disk array for enhanced perfor-mance and reliability. In Proc. IPDPS, April 2010, pp.1-12.
[14] Lee Y, Jung S, Song Y H. FRA: A flash-aware redundancy ar-ray of flash storage. In Proc. the 7th CODES+ISSS, October2009, pp.163-172.
[15] Chang Y B, Chang L P. A self-balancing striping scheme forNAND-flash storage systems. In Proc. the 2008 ACM Sym-posium on Applied Computing, March 2008, pp.1715-1719.
[16] Im S, Shin D K. Flash-aware RAID techniques for dependableand high-performance flash memory SSD. IEEE Transactionson Computers, 2011, 60(1): 80-92.
[17] Xiao N, Chen Z G, Liu F, Lai M C, An L F. P3Stor: Aparallel, durable flash-based SSD for enterprise-scale storagesystems. Science China Information Sciences, 2011, 54(6):1129-1141.
No related articles found!
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
[1] 张钹; 张恬; 张建伟; 张铃;. Motion Planning for Robots with Topological Dimension Reduction Method[J]. , 1990, 5(1): 1 -16 .
[2] 苏伯珙; 王剑; 夏金石;. TST——An Algorithm for Global Microcode Compaction with Timing Constraints[J]. , 1991, 6(1): 97 -107 .
[3] 眭跃飞;. The Polynomially Exponential Time Restrained Analytical Hierarchy[J]. , 1991, 6(3): 282 -284 .
[4] 马小虎; 潘志庚; 张福炎;. The Automatic Generation of Chinese Outline Font Based on Stroke Extraction[J]. , 1995, 10(1): 42 -52 .
[5] 高庆狮; 刘志勇;. K-Dimensional Optimal Parallel Algorithm for the Solution of a General Class of Recurrence Equations[J]. , 1995, 10(5): 417 -424 .
[6] 余盛可;. Reasoning in H-Net: A Unified Approach to Intelligent Hypermedia Systems[J]. , 1996, 11(1): 83 -89 .
[7] 郑方; 吴文虎; 方棣棠;. Center-Distance Continuous Probability Models and the Distance Measure[J]. , 1998, 13(5): 426 -437 .
[8] 傅育熙;. Reaction Graph[J]. , 1998, 13(6): 510 -530 .
[9] . 无线自主网中的链路分配最优化算法[J]. , 2006, 21(1): 89 -94 .
[10] . 颜色图像滤波的四元数扩散方法[J]. , 2006, 21(1): 126 -136 .
版权所有 © 《计算机科学技术学报》编辑部
本系统由北京玛格泰克科技发展有限公司设计开发 技术支持：support@magtech.com.cn
总访问量：