Parallel file systems stripe the data from a single file across multiple cluster/grid nodes so that the systems can access file in parallel. In such a system, if an I/O node or the storage device of that node doesn't work, all the subfiles on the node can't be accessed. In this paper, we introduce a special fault tolerance model for parallel file systems called Round-robin Redundant Backup of Subfile (RRBS), This model ensures the accessibility of the parallel files even when an I/O node is failure. In order to test the usability of RRBS, we also developed a prototype of parallel file system called WPFS on a PC/Windows cluster. © Springer-Verlag Berlin Heidelberg 2005.
CITATION STYLE
Huo, Y. M., Ju, J. B., & Hu, L. (2005). RRBS: A fault tolerance model for cluster/grid parallel file system. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3758 LNCS, pp. 180–187). https://doi.org/10.1007/11576235_23
Mendeley helps you to discover research relevant for your work.