The Community for Technology Leaders
2016 International Conference on Big Data and Smart Computing (BigComp) (2016)
Hong Kong, China
Jan. 18, 2016 to Jan. 20, 2016
ISSN: 2375-9356
ISBN: 978-1-4673-8795-8
pp: 386-389
Tae-Gun Song , Department of Electronic Engineering, Inha University, Incheon, South Korea
Mehdi Pirahandeh , Department of Electronic Engineering, Inha University, Incheon, South Korea
Deok-Hwan Kim , Department of Electronic Engineering, Inha University, Incheon, South Korea
ABSTRACT
Traditional hierarchical RAID causes huge GPU overhead and does not support node failure. To resolve this problem, this paper proposes a new hierarchical redundant array of inexpensive disks (RAID)'s parity generation using pass-through GPU in multi virtual-machine (VM) environment. The proposed method reduces GPU overhead and parity generation time, and supports node failure compared to the traditional hierarchical RAID. Experimental results show that the average total encoding performance of the proposed method using multiple VMs is improved by 50.11% compared to Khasymski's RAID 51 with respect to various file sizes.
INDEX TERMS
Graphics processing units, Encoding, Virtual machining, Kernel, Throughput, Cloud computing, Reliability
CITATION

Tae-Gun Song, M. Pirahandeh and Deok-Hwan Kim, "Hierarchical RAID's parity generation using pass-through GPU in multi virtual-machine environment," 2016 International Conference on Big Data and Smart Computing (BigComp)(BIGCOMP), Hong Kong, China, 2016, pp. 386-389.
doi:10.1109/BIGCOMP.2016.7425954
89 ms
(Ver 3.3 (11022016))