An Exponential Lower Bound on the Sub-Packetization of Minimum Storage Regenerating Codes

被引:8
作者
Alrabiah, Omar [1 ,2 ]
Guruswami, Venkatesan [2 ]
机构
[1] Carnegie Mellon Univ, Dept Math Sci, Pittsburgh, PA 15213 USA
[2] Carnegie Mellon Univ, Comp Sci Dept, Pittsburgh, PA 15213 USA
关键词
Distributed storage system (DSS); maximum distance separable (MDS) vector codes; repair bandwidth; sub-packetization; cutset bound; minimum storage regenerating (MSR) codes; OPTIMAL REPAIR; DISTRIBUTED STORAGE; MDS CODES; CONSTRUCTIONS; ACCESS; MSR;
D O I
10.1109/TIT.2021.3112286
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
An (n, k, l)-vector MDS code over a field F is a F-linear subspace of (F-l)(n) of dimension kl, such that any k (vector) symbols of the codeword suffice to determine the remaining r = n - k (vector) symbols. The length l of each codeword symbol is called the sub-packetization of the code. Such a code is called minimum storage regenerating (MSR), if any single symbol of a codeword can be recovered by downloading l/r field elements (which is known to be the minimum possible) from each of the other symbols. MSR codes are attractive for use in distributed storage systems, and by now a variety of ingenious constructions of MSR codes are available. However, they all suffer from exponentially large sub-packetization l greater than or similar to r(k/r). Our main result is an almost tight lower bound showing that for an MSR code, one must have l >= exp(Omega(k/r)). Previously, a lower bound of approximate to exp(root k/r), and a tight lower bound for a restricted class of "optimal access" MSR codes, were known.
引用
收藏
页码:8086 / 8093
页数:8
相关论文
共 26 条
  • [1] An Exponential Lower Bound on the Sub-Packetization of MSR Codes
    Alrabiah, Omar
    Guruswami, Venkatesan
    [J]. PROCEEDINGS OF THE 51ST ANNUAL ACM SIGACT SYMPOSIUM ON THEORY OF COMPUTING (STOC '19), 2019, : 979 - 985
  • [2] Balaji SB, 2018, IEEE INT SYMP INFO, P2381, DOI 10.1109/ISIT.2018.8437486
  • [3] Blaum M., 1994, Proceedings the 21st Annual International Symposium on Computer Architecture (Cat. No.94CH3397-7), P245, DOI 10.1109/ISCA.1994.288145
  • [4] Asymptotic Interference Alignment for Optimal Repair of MDS Codes in Distributed Storage
    Cadambe, Viveck R.
    Jafar, Syed Ali
    Maleki, Hamed
    Ramchandran, Kannan
    Suh, Changho
    [J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2013, 59 (05) : 2974 - 2987
  • [5] Cadambe VR, 2011, CONF REC ASILOMAR C, P1850, DOI 10.1109/ACSSC.2011.6190343
  • [6] A Survey on Network Codes for Distributed Storage
    Dimakis, Alexandros G.
    Ramchandran, Kannan
    Wu, Yunnan
    Suh, Changho
    [J]. PROCEEDINGS OF THE IEEE, 2011, 99 (03) : 476 - 489
  • [7] Network Coding for Distributed Storage Systems
    Dimakis, Alexandros G.
    Godfrey, P. Brighten
    Wu, Yunnan
    Wainwright, Martin J.
    Ramchandran, Kannan
    [J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2010, 56 (09) : 4539 - 4551
  • [8] On the Locality of Codeword Symbols
    Gopalan, Parikshit
    Huang, Cheng
    Simitci, Huseyin
    Yekhanin, Sergey
    [J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2012, 58 (11) : 6925 - 6934
  • [9] Minimum Storage Regenerating Codes for All Parameters
    Goparaju, Sreechakra
    Fazeli, Arman
    Vardy, Alexander
    [J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2017, 63 (10) : 6318 - 6328
  • [10] An Improved Sub-Packetization Bound for Minimum Storage Regenerating Codes
    Goparaju, Sreechakra
    Tamo, Itzhak
    Calderbank, Robert
    [J]. IEEE TRANSACTIONS ON INFORMATION THEORY, 2014, 60 (05) : 2770 - 2779