A proactive fault tolerance framework for high-performance computing

Antonina Litvinova; Christian Engelmann; Stephen L. Scott

doi:10.2316/p.2010.676-024

A proactive fault tolerance framework for high-performance computing

Antonina Litvinova, Christian Engelmann, Stephen L. Scott

Intelligent Systems and Facilities

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

13 Scopus citations

Abstract

As high-performance computing (HPC) systems continue to increase in scale, their mean-time to interrupt decreases respectively. The current state of practice for fault tolerance (FT) is checkpoint/restart. However, with increasing error rates, increasing aggregate memory and not proportionally increasing I/O capabilities, it is becoming less efficient. Proactive FT avoids experiencing failures through preventative measures, such as by migrating application parts away from nodes that are "about to fail". This paper presents a proactive FT framework that performs environmental monitoring, event logging, parallel job monitoring and resource monitoring to analyze HPC system reliability and to perform FT through such preventative actions.

Original language	English
Title of host publication	Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010
Publisher	Acta Press
Pages	105-110
Number of pages	6
ISBN (Print)	9780889868205
DOIs	https://doi.org/10.2316/p.2010.676-024
State	Published - 2010
Event	9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010 - Innsbruck, Austria Duration: Feb 16 2010 → Feb 18 2010

Publication series

Name	Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010

Conference

Conference	9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010
Country/Territory	Austria
City	Innsbruck
Period	02/16/10 → 02/18/10

Keywords

Fault tolerance
High availability
High-performance computing
Reliability
System monitoring

Access to Document

10.2316/p.2010.676-024

Cite this

Litvinova, A., Engelmann, C., & Scott, S. L. (2010). A proactive fault tolerance framework for high-performance computing. In Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010 (pp. 105-110). (Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010). Acta Press. https://doi.org/10.2316/p.2010.676-024

Litvinova, Antonina ; Engelmann, Christian ; Scott, Stephen L. / A proactive fault tolerance framework for high-performance computing. Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010. Acta Press, 2010. pp. 105-110 (Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010).

@inproceedings{773d4e2944cd4f378165da536abcb0a9,

title = "A proactive fault tolerance framework for high-performance computing",

abstract = "As high-performance computing (HPC) systems continue to increase in scale, their mean-time to interrupt decreases respectively. The current state of practice for fault tolerance (FT) is checkpoint/restart. However, with increasing error rates, increasing aggregate memory and not proportionally increasing I/O capabilities, it is becoming less efficient. Proactive FT avoids experiencing failures through preventative measures, such as by migrating application parts away from nodes that are {"}about to fail{"}. This paper presents a proactive FT framework that performs environmental monitoring, event logging, parallel job monitoring and resource monitoring to analyze HPC system reliability and to perform FT through such preventative actions.",

keywords = "Fault tolerance, High availability, High-performance computing, Reliability, System monitoring",

author = "Antonina Litvinova and Christian Engelmann and Scott, {Stephen L.}",

year = "2010",

doi = "10.2316/p.2010.676-024",

language = "English",

isbn = "9780889868205",

series = "Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010",

publisher = "Acta Press",

pages = "105--110",

booktitle = "Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010",

note = "9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010 ; Conference date: 16-02-2010 Through 18-02-2010",

}

Litvinova, A, Engelmann, C & Scott, SL 2010, A proactive fault tolerance framework for high-performance computing. in Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010. Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010, Acta Press, pp. 105-110, 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010, Innsbruck, Austria, 02/16/10. https://doi.org/10.2316/p.2010.676-024

A proactive fault tolerance framework for high-performance computing. / Litvinova, Antonina; Engelmann, Christian; Scott, Stephen L.
Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010. Acta Press, 2010. p. 105-110 (Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - A proactive fault tolerance framework for high-performance computing

AU - Litvinova, Antonina

AU - Engelmann, Christian

AU - Scott, Stephen L.

PY - 2010

Y1 - 2010

N2 - As high-performance computing (HPC) systems continue to increase in scale, their mean-time to interrupt decreases respectively. The current state of practice for fault tolerance (FT) is checkpoint/restart. However, with increasing error rates, increasing aggregate memory and not proportionally increasing I/O capabilities, it is becoming less efficient. Proactive FT avoids experiencing failures through preventative measures, such as by migrating application parts away from nodes that are "about to fail". This paper presents a proactive FT framework that performs environmental monitoring, event logging, parallel job monitoring and resource monitoring to analyze HPC system reliability and to perform FT through such preventative actions.

AB - As high-performance computing (HPC) systems continue to increase in scale, their mean-time to interrupt decreases respectively. The current state of practice for fault tolerance (FT) is checkpoint/restart. However, with increasing error rates, increasing aggregate memory and not proportionally increasing I/O capabilities, it is becoming less efficient. Proactive FT avoids experiencing failures through preventative measures, such as by migrating application parts away from nodes that are "about to fail". This paper presents a proactive FT framework that performs environmental monitoring, event logging, parallel job monitoring and resource monitoring to analyze HPC system reliability and to perform FT through such preventative actions.

KW - Fault tolerance

KW - High availability

KW - High-performance computing

KW - Reliability

KW - System monitoring

UR - http://www.scopus.com/inward/record.url?scp=77954589337&partnerID=8YFLogxK

U2 - 10.2316/p.2010.676-024

DO - 10.2316/p.2010.676-024

M3 - Conference contribution

AN - SCOPUS:77954589337

SN - 9780889868205

T3 - Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010

SP - 105

EP - 110

BT - Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010

PB - Acta Press

T2 - 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010

Y2 - 16 February 2010 through 18 February 2010

ER -

Litvinova A, Engelmann C, Scott SL. A proactive fault tolerance framework for high-performance computing. In Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010. Acta Press. 2010. p. 105-110. (Proceedings of the 9th IASTED International Conference on Parallel and Distributed Computing and Networks, PDCN 2010). doi: 10.2316/p.2010.676-024

A proactive fault tolerance framework for high-performance computing

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this