Effect of Codeword Placement on the Reliability of Erasure Coded Data Storage Systems

Modern data storage systems employ advanced erasure codes to protect data from storage node failures because of their ability to provide high data reliability at high storage efficiency. In contrast to previous studies, we consider the practical case where the length of codewords in an erasure coded system is much smaller than the number of storage nodes in the system. In this case, there exists a large number of possible ways in which different codewords can be stored across the nodes of the system. In this paper, it is shown that a declustered placement of codewords can significantly improve system reliability compared to other placement schemes. A detailed reliability analysis is presented that accounts for the rebuild times involved, the amounts of partially rebuilt data when additional nodes fail during rebuild, and an intelligent rebuild process that attempts to rebuild the most critical codewords first.

[1]  Peter Druschel,et al.  Peer-to-peer systems , 2010, Commun. ACM.

[2]  Alexandros G. Dimakis,et al.  Distributed Storage Allocation for High Reliability , 2010, 2010 IEEE International Conference on Communications.

[3]  Van-Anh Truong,et al.  Availability in Globally Distributed Storage Systems , 2010, OSDI.

[4]  Ilias Iliadis,et al.  A General Reliability Model for Data Storage Systems , 2012, 2012 Ninth International Conference on Quantitative Evaluation of Systems.

[5]  Jim Davies,et al.  A Comparison of Replication Strategies for Reliable Decentralised Storage , 2006, J. Networks.

[6]  Robert Haas,et al.  Reliability of Data Storage Systems under Network Rebuild Bandwidth Constraints , 2012, 2012 IEEE 20th International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunication Systems.

[7]  Mario Blaum,et al.  Mirrored Disk Organization Reliability Analysis , 2006, IEEE Transactions on Computers.

[8]  Mario Blaum,et al.  Higher reliability redundant disk arrays: Organization, operation, and coding , 2009, TOS.

[9]  Yunnan Wu,et al.  A Survey on Network Codes for Distributed Storage , 2010, Proceedings of the IEEE.

[10]  Ethan L. Miller,et al.  Reliability of flat XOR-based erasure codes on heterogeneous devices , 2008, 2008 IEEE International Conference on Dependable Systems and Networks With FTCS and DCC (DSN).

[11]  Garth A. Gibson,et al.  RAID: high-performance, reliable secondary storage , 1994, CSUR.

[12]  Xiaozhou Li,et al.  Reliability analysis of deduplicated and erasure-coded storage , 2011, PERV.

[13]  Randy H. Katz,et al.  A case for redundant arrays of inexpensive disks (RAID) , 1988, SIGMOD '88.

[14]  Joseph Pasquale,et al.  Analysis of Long-Running Replicated Systems , 2006, Proceedings IEEE INFOCOM 2006. 25TH IEEE International Conference on Computer Communications.

[15]  Ethan L. Miller,et al.  Evaluation of distributed recovery in large-scale storage systems , 2004, Proceedings. 13th IEEE International Symposium on High performance Distributed Computing, 2004..

[16]  John Kubiatowicz,et al.  Erasure Coding Vs. Replication: A Quantitative Comparison , 2002, IPTPS.

[17]  Christina Fragouli,et al.  Reliability of Clustered vs. Declustered Replica Placement in Data Storage Systems , 2011, 2011 IEEE 19th Annual International Symposium on Modelling, Analysis, and Simulation of Computer and Telecommunication Systems.

[18]  Hyokyung Bahn,et al.  P/PA-SPTF: Parallelism-aware request scheduling algorithms for MEMS-based storage devices , 2009, TOS.