簡易檢索 / 詳目顯示

研究生: 陳志誠
Chih-Cheng Chen
論文名稱: 使用parity cache 來減少小量寫在Linux software RAID-5上的負擔
Using parity cache to reduce overhead of small write requests under Linux software RAID-5 system
指導教授: 鍾葉青
Yeh-Ching Chung
口試委員:
學位類別: 碩士
Master
系所名稱: 電機資訊學院 - 資訊工程學系
Computer Science
論文出版年: 2008
畢業學年度: 96
語文別: 英文
論文頁數: 25
中文關鍵詞: 磁碟陣列磁碟快取磁碟緩衝儲存區輸入輸出Linux
外文關鍵詞: RAID, disk cache, disk buffer, storage, I/O, Linux
相關次數: 點閱:3下載:0
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報
  • Disk arrays, or RAIDs, have become the trend to increase the capacity and bandwidth of most storage system. Software RAID is the easy and economical way to activate a RAID system. RAID-5 is a common implementation taken by enterprise or business companies. However, the RAID-5 has small write issue, especially in the environments where small requests dominate the workload. There are many theses proposed their solutions to solve this point. In this paper, we also try to propose another method to enhance the performance of disk I/O on RAID-5. We use a part of memory as our cache to storage the frequent data on RAID-5, and we can reduce the frequency of access disk. The simulation results show that our design can improve the performance about 7%~12% regarding progress which need to frequently small write.


    磁碟陣列或稱RAID,增加儲存系統的容量和頻寬,已經成為發展的趨勢。軟件RAID是一種簡單而且經濟的辦法去啟動一個RAID的系統。RAID-5 是一種企業或商業公司常見的實作方式。然而尤其在小量寫支配工作量的環境□,RAID-5有小量寫的問題。有很多論文對於這點提出他們的解決之道。在本篇論文中,我們也試著去提出另一種方法去加強RAID-5上面硬碟輸入輸出的性能。我們使用記憶體當作我們的快速儲存區去儲存RAID-5上面經常出現的資料,這樣一來我們可以減少它存取硬碟的量。實驗的結果顯示對於那些經常需要去小量寫的程式,我們的設計可以增進大約7%~12%的效能。

    Chapter 1 Introduction 1 Chapter 2 Related Work 3 Chapter 3 Linux Software RAID-5 Architecture 5 3.1 The I/O background 5 3.1.1 The flow chart of I/O operation of Linux 5 3.1.2 The page cache and direct I/O 6 3.1.3 The BIO structure 7 3.2 RAID-5 module 8 3.2.1 Overview 8 3.2.2 Stripe 9 3.2.3 Rmw (read_modify_write) v.s. rcw (read_reconstruct_write) 10 Chapter 4 The Implementation of Advanced Parity Cache (APC) 12 4.1 Overview 12 4.2 Framework 13 4.3 The stored information 14 4.4 The scenario of our APC 15 Chapter 5 Experimental Results 20 5.1 Platform 20 5.2 Small write on RAID-5 20 5.3 The results 21 Chapter 6 Conclusions 24 References 25

    [1] Varma, A.; Jacobson, Q., "Destage algorithms for disk arrays with nonvolatile caches," Computers, IEEE Transactions on , vol.47, no.2, pp.228-235, Feb 1998
    [2] Yiming Hu; Nightingale, T.; Qing Yang, "RAPID-Cache-A reliable and inexpensive write cache for high performance storage systems," Parallel and Distributed Systems, IEEE Transactions on , vol.13, no.3, pp.290-307, Mar 2002
    [3] Jun Wang; Yiming Hu, "A novel reordering write buffer to improve write performance of log-structured file systems," Computers, IEEE Transactions on , vol.52, no.12, pp. 1559-1572, Dec. 2003
    [4] Menon, J., "A performance comparison of RAID-5 and log-structured arrays," High Performance Distributed Computing, 1995., Proceedings of the Fourth IEEE International Symposium on , vol., no., pp.167-178, 2-4 Aug 1995
    [5] J. and Douglis, F. 1989. Beating the I/O bottleneck: a case for log-structured file systems. SIGOPS Oper. Syst. Rev. 23, 1 (Jan. 1989), 11-28.
    [6] Peng Cheng; Hai Jin; Xinrong Zhou; Qiong Chen; Jiangling Zhang, "HUST RAID: high performance RAID in real-time system," Communications, Computers and Signal Processing, 1999 IEEE Pacific Rim Conference on , vol., no., pp.59-62, 1999
    [7] Hou, R.Y.; Menon, J.; Patt, Y.N., "Balancing I/O response time and disk rebuild time in a RAID5 disk array," System Sciences, 1993, Proceeding of the Twenty-Sixth Hawaii International Conference on , vol.i, no., pp. 70-79 vol.1, 5-8 Jan 1993
    [8] Thomasian, A.; Jai Menon, "RAID5 performance with distributed sparing," Parallel and Distributed Systems, IEEE Transactions on , vol.8, no.6, pp.640-657, Jun 1997

    無法下載圖示 全文公開日期 本全文未授權公開 (校內網路)
    全文公開日期 本全文未授權公開 (校外網路)

    QR CODE