On safety and timeliness in distributed data management (preliminary report)

Danny Dolev; H. Raymond Strong

doi:10.1007/3-540-50403-6_32

On safety and timeliness in distributed data management (preliminary report)

Danny Dolev, H. Raymond Strong

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

1 Scopus citations

Abstract

When partitions may occur with a positive probability there is no way to guarantee safety and timely operation. In most cases the probability of partition is small even though some components may be inoperative for long periods of time. Blocking prevents transient faults from causing a loss of consistency in the rare instances of partition. Much as we would like a truly nonblocking coordination protocol, Theorem 3.2 implies that guaranteed eventual communication is not enough to guarantee safety without blocking in an asynchronous environment, and Corollary 4.2 says that even in a synchronous environment, any safe atomic commit protocol is blocking.

Original language	English
Title of host publication	CONCURRENCY 1988 - International Conference on Concurrency, Proceedings
Editors	Frederich H. Vogt
Publisher	Springer Verlag
Pages	64-72
Number of pages	9
ISBN (Print)	9783540504030
DOIs	https://doi.org/10.1007/3-540-50403-6_32
State	Published - 1988
Externally published	Yes
Event	International Conference on Concurrency Theory, CONCURRENCY 1988 - Hamburg, Germany Duration: 18 Oct 1988 → 19 Oct 1988

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	335 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	International Conference on Concurrency Theory, CONCURRENCY 1988
Country/Territory	Germany
City	Hamburg
Period	18/10/88 → 19/10/88

Bibliographical note

Publisher Copyright:
© 1988, Springer-Verlag.

Access to Document

10.1007/3-540-50403-6_32

Cite this

Dolev, D., & Raymond Strong, H. (1988). On safety and timeliness in distributed data management (preliminary report). In F. H. Vogt (Ed.), CONCURRENCY 1988 - International Conference on Concurrency, Proceedings (pp. 64-72). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 335 LNCS). Springer Verlag. https://doi.org/10.1007/3-540-50403-6_32

Dolev, Danny ; Raymond Strong, H. / On safety and timeliness in distributed data management (preliminary report). CONCURRENCY 1988 - International Conference on Concurrency, Proceedings. editor / Frederich H. Vogt. Springer Verlag, 1988. pp. 64-72 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{36556e9dfdbd4c95b613cea93321bec3,

title = "On safety and timeliness in distributed data management (preliminary report)",

abstract = "When partitions may occur with a positive probability there is no way to guarantee safety and timely operation. In most cases the probability of partition is small even though some components may be inoperative for long periods of time. Blocking prevents transient faults from causing a loss of consistency in the rare instances of partition. Much as we would like a truly nonblocking coordination protocol, Theorem 3.2 implies that guaranteed eventual communication is not enough to guarantee safety without blocking in an asynchronous environment, and Corollary 4.2 says that even in a synchronous environment, any safe atomic commit protocol is blocking.",

author = "Danny Dolev and {Raymond Strong}, H.",

note = "Publisher Copyright: {\textcopyright} 1988, Springer-Verlag.; International Conference on Concurrency Theory, CONCURRENCY 1988 ; Conference date: 18-10-1988 Through 19-10-1988",

year = "1988",

doi = "10.1007/3-540-50403-6_32",

language = "אנגלית",

isbn = "9783540504030",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "64--72",

editor = "Vogt, {Frederich H.}",

booktitle = "CONCURRENCY 1988 - International Conference on Concurrency, Proceedings",

address = "גרמניה",

}

Dolev, D & Raymond Strong, H 1988, On safety and timeliness in distributed data management (preliminary report). in FH Vogt (ed.), CONCURRENCY 1988 - International Conference on Concurrency, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 335 LNCS, Springer Verlag, pp. 64-72, International Conference on Concurrency Theory, CONCURRENCY 1988, Hamburg, Germany, 18/10/88. https://doi.org/10.1007/3-540-50403-6_32

On safety and timeliness in distributed data management (preliminary report). / Dolev, Danny; Raymond Strong, H.
CONCURRENCY 1988 - International Conference on Concurrency, Proceedings. ed. / Frederich H. Vogt. Springer Verlag, 1988. p. 64-72 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 335 LNCS).

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

TY - GEN

T1 - On safety and timeliness in distributed data management (preliminary report)

AU - Dolev, Danny

AU - Raymond Strong, H.

PY - 1988

Y1 - 1988

N2 - When partitions may occur with a positive probability there is no way to guarantee safety and timely operation. In most cases the probability of partition is small even though some components may be inoperative for long periods of time. Blocking prevents transient faults from causing a loss of consistency in the rare instances of partition. Much as we would like a truly nonblocking coordination protocol, Theorem 3.2 implies that guaranteed eventual communication is not enough to guarantee safety without blocking in an asynchronous environment, and Corollary 4.2 says that even in a synchronous environment, any safe atomic commit protocol is blocking.

AB - When partitions may occur with a positive probability there is no way to guarantee safety and timely operation. In most cases the probability of partition is small even though some components may be inoperative for long periods of time. Blocking prevents transient faults from causing a loss of consistency in the rare instances of partition. Much as we would like a truly nonblocking coordination protocol, Theorem 3.2 implies that guaranteed eventual communication is not enough to guarantee safety without blocking in an asynchronous environment, and Corollary 4.2 says that even in a synchronous environment, any safe atomic commit protocol is blocking.

UR - http://www.scopus.com/inward/record.url?scp=84936662929&partnerID=8YFLogxK

U2 - 10.1007/3-540-50403-6_32

DO - 10.1007/3-540-50403-6_32

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:84936662929

SN - 9783540504030

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 64

EP - 72

BT - CONCURRENCY 1988 - International Conference on Concurrency, Proceedings

A2 - Vogt, Frederich H.

PB - Springer Verlag

T2 - International Conference on Concurrency Theory, CONCURRENCY 1988

Y2 - 18 October 1988 through 19 October 1988

ER -

Dolev D, Raymond Strong H. On safety and timeliness in distributed data management (preliminary report). In Vogt FH, editor, CONCURRENCY 1988 - International Conference on Concurrency, Proceedings. Springer Verlag. 1988. p. 64-72. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/3-540-50403-6_32

On safety and timeliness in distributed data management (preliminary report)

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this