doi:

DOI: 10.3724/SP.J.1001.2012.04046

Journal of Software (软件学报) 2012/23:4 PP.987-995

DMStone: A Tool for Evaluating Hierarchical Storage Management Systems


Abstract:
Realistic system states and representative workloads are required to evaluate a hierarchical storage management (HSM) system. The existing method to evaluate a HSM system reconstructs system states by replaying a trace within period of time. Ignoring files not used recently, the result of this method is not convincing enough. This paper proposes DMStone, a tool for evaluating HSM systems. DMStone uses file-system snapshots to generate a system state. Furthermore, it extracts workload characteristics by comparing two contiguous snapshots and generates representative workloads. DMStone can provide a realistic file-system state, including files used recently and those not used for a long time. Moreover, it makes subsequent workloads conform to realistic system states.

Key words:hierarchical storage,performance evaluation,system state,I/O workload,file system snapshot

ReleaseDate:2014-07-21 16:14:32



[1] Gibson TJ, Miller EL, Long DDE. Long-Term file activity and inter-reference patterns. In: Proc. of the 24th Int’l Conf. on Technology Management and Performance Evaluation of Enterprise-Wide Information Systems. Anaheim, 1998. 976-987.

[2] Smith AJ. Long term file migration: Development and evaluation of algorithms. Communications of the ACM, 1981,24(8):521-532. [doi: 10.1145/358722.358737]

[3] Salmon E, Tarshish A, Palm N, Patel S, Saletta M, Vanderlan E, Rouch M, Burns L, Duffy DD, Caine R, Golay R, Paffel J, Schumann N. Hierarchical storage management at the NASA center for computational sciences: From unitree to SAM-QFS. In: Kobler B, Hariharan PC, eds. Proc. of the 12th NASA Goddard, 21st IEEE Conf. on Mass Storage Systems and Technologies (MSST 2004). Washington: IEEE, 2004. 177-183.

[4] Baker MG, Hartman JH, Kupfer MD, Shirriff KW, Ousterhout JK. Measurements of a distributed file system. In: Levy HM, ed. Proc. of the 13th ACM Symp. on Operating Systems Principles. New York: ACM, 1991. 198-212. [doi: 10.1145/121132.121164]

[5] Gribble SD, Manku GS, Roselli DS, Brewer EA, Gibson TJ, Miller EL. Self-Similarity in file systems. In: Leutenegger S, ed. Proc. of the 1998 ACM SIGMETRICS Joint Int’l Conf. on Measurement and Modeling of Computer Systems. New York: ACM, 1998. 141-150. [doi: 10.1145/277851.277894]

[6] Ousterhout JK, Costa HD, Harrison D, Kunze JA, Kupfer M, Thompson JG. A trace-driven analysis of the UNIX 4.2 BSD files system. In: Waite WM, ed. Proc. of the 10th ACM Symp. on Operating Systems Principles. New York: ACM, 1985. 15-24. [doi: 10.1145/323647.323631]

[7] Riedel E, Kallahalla M, Swaminathan R. A framework for evaluating storage system security. In: Proc. of the FAST 2002 Conf. on File and Storage Technologies. Berkeley: USENIX Association, 2002. 15-30.

[8] SNIA. Storage network industry association: Iotta repository. 2007. http://iotta.snia.org

[9] Anderson DC, Chase JS. Fstress: A flexible network file service benchmark. Technical Report, Duke University, 2002.

[10] Ebling MR, Satyanarayanan M. Synrgen: An extensible file reference generator. In: Gaither BD, ed. Proc. of the 1994 ACM SIGMETRICS Conf. on Measurement and Modeling of Computer Systems. New York: ACM, 1994. 108-117. [doi: 10.1145/ 183018.183030]

[11] Anderson E, Kallahalla M, Uysal M, Swaminathan R. Buttress: A toolkit for flexible and high fidelity I/O benchmarking. In: Proc. of the FAST 2004 Conf. on File and Storage Technologies. San Jose: USENIX Association, 2004. 45-58.

[12] Mesnier MP, Wachs M, Sambasivan RR, Lopez J, Hendricks J, Ganger GR, O’Hallaron D. Trace: Parallel trace replay with approximate causal events. In: Proc. of the 5th USENIX Conf. on File and Storage Technologies. Berkeley: USENIX Association, 2007. 153-167.

[13] Traeger A, Zadok E, Joukov N, Wright CP. A nine year study of file system and storage benchmarking. ACM Trans. on Storage (TOS), 2008,4(2):1-56. [doi: 10.1145/1367829.1367831]

[14] Agrawal N, Arpaci-Dusseau AC, Arpaci-Dusseau RH. Generating realistic impressions for file-system benchmarking. In: Seltzer MI, Wheeler R, eds. Proc. of the 7th Conf. on File and Storage Technologies. Berkeley: USENIX Association, 2009. 125-138. [doi: 10.1145/1629080.1629086]

[15] OSDL. Iometer project. 2004. http://www.iometer.org/

[16] Floyd R. Short-Term file reference patterns in a Unix environment. 2009. https://urresearch.rochester.edu/institutionalPublication PublicView.action?institutionalItemId=6362

[17] Pike R, Presotto D, Thompson K, Trickey H. Plan 9 from Bell Labs. Computing Systems, 1995,8(3):221-254.

[18] Wang F, Xin Q, Hong B, Brandt SA, Miller EL, Long DDE, Mclarty TT. File system workload analysis for large scale scientific computing applications. In: Kobler B, Hariharan PC, eds. Proc. of the 21st IEEE Conf. on Mass Storage Systems and Technologies/ 12th NASA Goddard Conf. on Mass Storage Systems and Technologie. Washingtong: IEEE, 2004. 139-152.