The Journal of Korean Institute of Communications and Information Sciences (한국통신학회논문지)

The Korean Institute of Commucations and Information Sciences (한국통신학회)
권호 표지
DOI QR코드

DOI QR Code

Implementation and Performance Measuring of Erasure Coding of Distributed File System

분산 파일시스템의 소거 코딩 구현 및 성능 비교

  • Kim, Cheiyol (Electronics and Telecommunications Research Institute) ;
  • Kim, Youngchul (Electronics and Telecommunications Research Institute) ;
  • Kim, Dongoh (Electronics and Telecommunications Research Institute) ;
  • Kim, Hongyeon (Electronics and Telecommunications Research Institute) ;
  • Kim, Youngkyun (Electronics and Telecommunications Research Institute) ;
  • Seo, Daewha (Dept. of Electronics Eng., Kyungpook National University)
  • Received : 2016.10.10
  • Accepted : 2016.10.31
  • Published : 2016.11.30
Download PDF
⟨ Previous Next ⟩

Abstract

With the growth of big data, machine learning, and cloud computing, the importance of storage that can store large amounts of unstructured data is growing recently. So the commodity hardware based distributed file systems such as MAHA-FS, GlusterFS, and Ceph file system have received a lot of attention because of their scale-out and low-cost property. For the data fault tolerance, most of these file systems uses replication in the beginning. But as storage size is growing to tens or hundreds of petabytes, the low space efficiency of the replication has been considered as a problem. This paper applied erasure coding data fault tolerance policy to MAHA-FS for high space efficiency and introduces VDelta technique to solve data consistency problem. In this paper, we compares the performance of two file systems, MAHA-FS and GlusterFS. They have different IO processing architecture, the former is server centric and the latter is client centric architecture. We found the erasure coding performance of MAHA-FS is better than GlusterFS.

최근의 빅데이터, 머신러닝, 클라우드 컴퓨팅 분야의 성장에 따라 대용량의 비정형 데이터를 저장할 수 있는 스토리지의 중요성은 날로 커지고 있다. 이에 따라 MAHA-FS, GlusterFS, Ceph 등의 개방형 하드웨어 기반의 분산 파일시스템 기술이 많은 주목을 받고 있다. 이러한 저비용 분산 파일시스템들은 데이터의 내결함성을 보장하기 위하여 초기에 복제 방식을 사용하였으나, 스토리지의 용량이 커질수록 복제 방식이 가지는 스토리지 공간의 저효율성이 점차 부각되면서 이를 보완하려는 방향으로 연구가 진행되고 있다. 본 논문은 복제방식을 대체하여 스토리지 공간 효율성을 향상시킬 수 있는 소거코딩 기법을 MAHA-FS 분산 파일시스템에 적용하여 스토리지의 효율성을 높이고, 소거코딩 지원에 따라 발생하는 데이터 일관성 문제를 해결하는 효율적인 방식으로 VDelta 기법을 제안하고 적용하였다. 본 논문은 MAHA-FS와 GlusterFS의 소거코딩의 구조적 차이점을 기술하고 두 파일시스템의 성능을 비교하여 MAHA-FS의 소거코딩 성능이 GlusterFS에 비해 우수함을 확인하였다.

Keywords

References

  1. S. Ghemawat, H. Gobioff, and S.-T. Leung, "The Google File System," in Proc. ACM SOSP, pp. 29-43, 2003.
  2. GlusterFS. http://www.gluster.com/ (accessed Sept. 2016).
  3. S. A. Weil, S. A. Brandt, E. L. Miller, D. D. Long, and C. Maltzahn, "Ceph: A scalable, high-performance distributed file system," in Proc. 7th Symp. Operating Syst. Design and Implementation, USENIX Association, 2006.
  4. K. Shvachko, H. Kuang, S. Radia, and R. Chansler, "The hadoop distributed file system," IEEE MSST, pp. 1-10, 2010.
  5. H. Y. Kim, G. S. Jin, M. H. Cha, S. M. Lee, S. M. Lee, Y. C. Kim, and Y. K. Kim, "GLORY-FS: A distributed file system for large-scale internet service," KICS Inf. and Commun. Mag., vol. 30, no. 4, pp. 16-22, Mar. 2013.
  6. Y. C. Kim, D. O. Kim, H. Y. Kim, Y. K. Kim, and W. Choi, "MAHA-FS: A distributed file system for high performance metadata processing and random IO," KIPS Trans. Software and Data Eng., vol. 2, no. 2, pp. 91-96, Feb. 2013. https://doi.org/10.3745/KTSDE.2013.2.2.091
  7. J. S. Kim and T. W. Kim, "OwFS: A distributed file system for large-scale internet services," J. Korean Data & Inf. Sci. Soc., vol. 27, no. 5, pp. 77-85, May 2009.
  8. G. J. Lee, Y. C. Shin, J. H. Koo, and S. H. Choi, "Practical implementation and performance evaluation of random linear network coding," J. KICS, vol. 40, no. 9, pp. 1786-1792, Sept. 2015. https://doi.org/10.7840/kics.2015.40.9.1786
  9. D. Lambright, Erasure Codes and Storage Tiers on Gluster, SA summit, Sept. 23, 2014.
  10. A. Ajisaka, HDFS 2015: Past, Present, and Future(2015), Retrieved Sep., 30, 2016, from http://events.linuxfoundation.org/sites/events/files/slides/HDFS2015_Past_present_future.pdf
  11. T. Y. Kim, Lessons learned from deploying SSD in NAVER services(2014), Retrieved Sep., 30, 2016, http://dcslab.hanyang.ac.kr/nvramos/nvramos14/presentation/s1.pdf
  12. J. N. Gray, "Notes on data base operating systems," Springer-Verlag, vol. 60, pp. 393-481, Berlin, 1978.
  13. C. Gray and D. Cheriton, "Leases: An efficient fault-tolerant mechanism for distributed file cache consistency," ACM SIGOPS Operating Systems Rev., vol. 23, no. 5, pp. 202-210, Dec. 1989.
  14. H.-T. Kung and John T. Robinson, "On optimistic methods for concurrency control," ACM TODS, vol. 6, no. 2, pp. 213-226, Jun. 1981. https://doi.org/10.1145/319566.319567
  15. David P. Reed and L. Svobodova, "SWALLOW: A distributed data storage system for a local network," in Proc. 1FIP Working Group 6.4 Int. Workshop on Local Netw., pp. 355-373, Aug. 1980.
  16. David P. Reed, "Implementing atomic actions on decentralized data," ACM Trans. Computer Systems (TOCS), vol. 1, no. 1, pp. 3-23, Feb. 1983. https://doi.org/10.1145/357353.357355
  17. G. R. Garth, J. J. Wylie, R. G. Ganger, and M. K. Reiter, "Efficient consistency for erasure-coded data via versioning servers," Carnegie-Mellon Univ. Pittsburgh PA School of Comput. Sci., no. CMU-CS-03-127, Mar. 2003.
  18. S. Frolund, A. Merchant, Y. Saito, S. Spence, and A. Veitch, "A decentralized algorithm for erasure-coded virtual disks," IEEE Int. Conf. Dependable Syst. and Netw., pp. 125-134, Jun. 2004.
  19. J. S. Plank, S. Simmerman, and C. D. Schuman, Jerasure: A library in C/C++ facilitating erasure coding for storage applications-Version 1.2, Technical Report CS-08-627, University of Tennessee, 2008.
  20. Intel Storage Acceleration Library, https://software.intel.com/en-us/storage/ISA-L (accessed Sept. 2016).
  21. Iozone Filesystem Benchmark, http://www.iozone.org/docs/IOzone_msword_98.pdf(access ed Sept. 2016).
  22. "Filesystem in Userspace," https://sourceforge.net/projects/fuse/(accessed May 2016).
  23. D. H. Kim and S. Y. Hwang, "An efficient wear-leveling algorithm for NAND flash SSD with multi-channel and multi-way architecture," J. KICS, vol. 39B, no. 7, pp. 425-432, Jul. 2014. https://doi.org/10.7840/kics.2014.39B.7.425
  24. S. M. Han, H. S. Park, and T. W. Kwon. "Shelf-life time based cache replacement policy suitable for web environment," J. KICS, vol. 40, no. 6, pp. 1091-1101, Jun. 2015. https://doi.org/10.7840/kics.2015.40.6.1091