GPFS: A Shared-Disk File System for Large Computing Clusters

DMCA

GPFS: A Shared-Disk File System for Large Computing Clusters (2002)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Frank Schmuck , Roger Haskin

Venue:	In Proceedings of the 2002 Conference on File and Storage Technologies (FAST
Citations:	514 - 3 self

BibTeX

@INPROCEEDINGS{Schmuck02gpfs:a,
    author = {Frank Schmuck and Roger Haskin},
    title = {GPFS: A Shared-Disk File System for Large Computing Clusters},
    booktitle = {In Proceedings of the 2002 Conference on File and Storage Technologies (FAST},
    year = {2002},
    pages = {231--244}
}

OpenURL

Abstract

GPFS is IBM's parallel, shared-disk file system for cluster computers, available on the RS/6000 SP parallel supercomputer and on Linux clusters. GPFS is used on many of the largest supercomputers in the world. GPFS was built on many of the ideas that were developed in the academic community over the last several years, particularly distributed locking and recovery technology. To date it has been a matter of conjecture how well these ideas scale. We have had the opportunity to test those limits in the context of a product that runs on the largest systems in existence. While in many cases existing ideas scaled well, new approaches were necessary in many key areas. This paper describes GPFS, and discusses how distributed locking and recovery techniques were extended to scale to large clusters.

Keyphrases

shared-disk file system large computing cluster linux cluster many case distributed locking academic community sp parallel supercomputer recovery technology recovery technique large cluster new approach cluster computer many key area idea scale last several year