Author:
Yuan Wei,Yu Ying,Gao Yan,Tang Dan
Abstract
Abstract
The array code has the advantages of high computational efficiency and simple structure. And the erasure code strategy based on the array code is one of the ideal fault-tolerant methods for distributed storage systems. However, low fault tolerance has always been a major obstacle to the practicality of array code. According to the known literature, the fault-tolerance of MDS array code is no more than three. And the array code with the biggest fault tolerance is the Grid code. Its fault tolerance is usually only 15. In view of this situation, this paper proposes a kind of multi-fault-tolerant MDS array code, called Reed-Solomon Array (RA). The RA code has higher coding efficiency, whose encoding process and decoding process only involve binary XOR operations. What is more, the construction method is simple and easy to be implemented by software and hardware. And then the RA code is a MDS code with optimal storage efficiency. Finally the size of the storage array of the RA code is not strictly limited. The extensibility of the distributed storage system is met.
Subject
General Physics and Astronomy
Reference17 articles.
1. The Google file system[J];Ghemawat;ACM SIGOPS Operating Systems Review,2003
2. The Hadoop Distributed File System[C] / / IEEE, Symposium on MASS Storage Systems and Technologies;Shvachko,2010
3. Sector-Disk (SD) Erasure Codes for Mixed Failure Modes in RAID Systems[J];Plank;Acm Transactions on Storage,2013
4. EVENODD: an optimal scheme for tolerating double disk failures in RAID architectures[P];Blaum,1994