• DocumentCode
    3118167
  • Title

    Reliability Analysis for Full-2 Code

  • Author

    Lin, Sheng ; Zhang, Chi ; Wang, Gang ; Liu, Xiaoguang ; Liu, Jing

  • Author_Institution
    Nankai-Baidu Joint Lab., Nankai Univ., Tianjin, China
  • fYear
    2009
  • fDate
    14-16 Dec. 2009
  • Firstpage
    454
  • Lastpage
    459
  • Abstract
    Recently, with the fast development of storage system, 2-erasure coding schemes were widely used in industrial society. To meet different requirements, many kinds of 2-erasure coding schemes were presented, such as Reed-Solomon codes, binary linear codes, parity array codes, and so on. Full-2 code is a 2-erasure binary linear code. It is a non-MDS code, but achieves optimal encoding, decoding, and updating performance. Moreover, its fault tolerance is beyond 2, i.e. ¿2-erasure¿ is the huge undervaluation of its fault tolerance. It is hard to evaluate the precise reliability of full-2 code. The reason is that the reliability model is complex and the proportion of recoverable k-erasures (k > 2) to total k-erasures is difficult to calculate. In this paper, we present a combinatorial method to analyze the precise reliability of full-2 code. The reliability of full-2 based storage systems is also evaluated.
  • Keywords
    binary codes; combinatorial mathematics; decoding; fault tolerance; linear codes; 2-erasure binary linear code; combinatorial method; decoding; fault tolerance; full-2 code; nonMDS code; optimal encoding; recoverable k-erasures; reliability analysis; storage system; Algorithm design and analysis; Computational complexity; Educational institutions; Encoding; Fault tolerance; Hard disks; Information analysis; Information technology; Linear code; Reed-Solomon codes; erasure code; full-2 code; graph representation; markov model; reliability;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Pervasive Systems, Algorithms, and Networks (ISPAN), 2009 10th International Symposium on
  • Conference_Location
    Kaohsiung
  • Print_ISBN
    978-1-4244-5403-7
  • Type

    conf

  • DOI
    10.1109/I-SPAN.2009.95
  • Filename
    5381581