Stable Checkpointing in Distributed Systems without Shared Disks

Peter Sobe

doi:10.1109/IPDPS.2003.1213392

Stable Checkpointing in Distributed Systems without Shared Disks

Peter Sobe

Institute of Computer Engineering

Abstract

Interacting processes an distributed systems save their checkpoints on local disks for efficiency reasons. But, because local checkpoints get unavailable with failing hosts, redundancy schemes similar to RAID-like storage schemes have to be used. In such systems, checkpoints are stable under a particular fault model because they can get reconstructed in the distributed system. In this paper, two variants of stable checkpoint storage are compared, (a) parity grouping over local checkpoints and (ii) RAID-like distribution of each checkpoint using a software based distributed storage system. An analysis is given to compare costs for collective checkpoint creation, recovery of a single process and rollback of all processes. The results show that despite the differences in detail, checkpointing using a distributed storage system is a reasonable solution.

Original language	English
Title of host publication	Proceedings International Parallel and Distributed Processing Symposium
Publisher	IEEE
Publication date	01.01.2003
Article number	1213392
ISBN (Print)	0-7695-1926-1
DOIs	https://doi.org/10.1109/IPDPS.2003.1213392
Publication status	Published - 01.01.2003
Event	International Parallel and Distributed Processing Symposium - Nice, France Duration: 22.04.2003 → 26.04.2003 Conference number: 115724

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/IPDPS.2003.1213392

Cite this

@inproceedings{afd2e6059be04e348a49688702fcb269,

title = "Stable Checkpointing in Distributed Systems without Shared Disks",

abstract = "Interacting processes an distributed systems save their checkpoints on local disks for efficiency reasons. But, because local checkpoints get unavailable with failing hosts, redundancy schemes similar to RAID-like storage schemes have to be used. In such systems, checkpoints are stable under a particular fault model because they can get reconstructed in the distributed system. In this paper, two variants of stable checkpoint storage are compared, (a) parity grouping over local checkpoints and (ii) RAID-like distribution of each checkpoint using a software based distributed storage system. An analysis is given to compare costs for collective checkpoint creation, recovery of a single process and rollback of all processes. The results show that despite the differences in detail, checkpointing using a distributed storage system is a reasonable solution.",

author = "Peter Sobe",

year = "2003",

month = jan,

day = "1",

doi = "10.1109/IPDPS.2003.1213392",

language = "English",

isbn = " 0-7695-1926-1",

series = "Proceedings - International Parallel and Distributed Processing Symposium",

publisher = "IEEE",

booktitle = "Proceedings International Parallel and Distributed Processing Symposium",

note = "International Parallel and Distributed Processing Symposium , IPDPS 2003 ; Conference date: 22-04-2003 Through 26-04-2003",

}

Sobe, P 2003, Stable Checkpointing in Distributed Systems without Shared Disks. in Proceedings International Parallel and Distributed Processing Symposium., 1213392, Proceedings - International Parallel and Distributed Processing Symposium, IEEE, International Parallel and Distributed Processing Symposium , Nice, France, 22.04.03. https://doi.org/10.1109/IPDPS.2003.1213392

TY - GEN

T1 - Stable Checkpointing in Distributed Systems without Shared Disks

AU - Sobe, Peter

N1 - Conference code: 115724

PY - 2003/1/1

Y1 - 2003/1/1

N2 - Interacting processes an distributed systems save their checkpoints on local disks for efficiency reasons. But, because local checkpoints get unavailable with failing hosts, redundancy schemes similar to RAID-like storage schemes have to be used. In such systems, checkpoints are stable under a particular fault model because they can get reconstructed in the distributed system. In this paper, two variants of stable checkpoint storage are compared, (a) parity grouping over local checkpoints and (ii) RAID-like distribution of each checkpoint using a software based distributed storage system. An analysis is given to compare costs for collective checkpoint creation, recovery of a single process and rollback of all processes. The results show that despite the differences in detail, checkpointing using a distributed storage system is a reasonable solution.

AB - Interacting processes an distributed systems save their checkpoints on local disks for efficiency reasons. But, because local checkpoints get unavailable with failing hosts, redundancy schemes similar to RAID-like storage schemes have to be used. In such systems, checkpoints are stable under a particular fault model because they can get reconstructed in the distributed system. In this paper, two variants of stable checkpoint storage are compared, (a) parity grouping over local checkpoints and (ii) RAID-like distribution of each checkpoint using a software based distributed storage system. An analysis is given to compare costs for collective checkpoint creation, recovery of a single process and rollback of all processes. The results show that despite the differences in detail, checkpointing using a distributed storage system is a reasonable solution.

UR - https://www.scopus.com/record/display.uri?eid=2-s2.0-84947297469&origin=inward&txGid=8a0f07de6648f248a41aebd8b2f6387f

U2 - 10.1109/IPDPS.2003.1213392

DO - 10.1109/IPDPS.2003.1213392

M3 - Conference contribution

SN - 0-7695-1926-1

T3 - Proceedings - International Parallel and Distributed Processing Symposium

BT - Proceedings International Parallel and Distributed Processing Symposium

PB - IEEE

T2 - International Parallel and Distributed Processing Symposium

Y2 - 22 April 2003 through 26 April 2003

ER -

Stable Checkpointing in Distributed Systems without Shared Disks

Abstract

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this