Leader election for replicated services using application scores

Diogo Becker, Flavio Junqueira, Marco Serafini

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Replicated services often rely on a leader to order client requests and broadcast state updates. In this work, we present POLE, a leader election algorithm that select leaders using application-specific scores. This flexibility given to the application enables the algorithm to tailor leader election according to metrics that are relevant in practical settings and that have been overlooked by existing approaches. Recovery time and request latency are examples of such metrics. To evaluate POLE, we use ZooKeeper, an open-source replicated service used for coordinating Web-scale applications. Our evaluation over realistic wide-area settings shows that application scores can have a significant impact on performance, and that just optimizing the latency of consensus does not translate into lower latency for clients. An important conclusion from our results is that obtaining a general strategy that satisfies a wide range of requirements is difficult, which implies that configurability is indispensable for practical leader election.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages289-308
Number of pages20
Volume7049 LNCS
DOIs
Publication statusPublished - 23 Dec 2011
Externally publishedYes
Event12th ACM/IFIP/USENIX International Middleware Conference, Middleware 2011 - Lisbon, Portugal
Duration: 12 Dec 201116 Dec 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7049 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other12th ACM/IFIP/USENIX International Middleware Conference, Middleware 2011
CountryPortugal
CityLisbon
Period12/12/1116/12/11

Fingerprint

Leader Election
Latency
Metric
Open Source
Broadcast
Recovery
Update
Flexibility
Imply
Evaluate
Requirements
Evaluation
Range of data

Keywords

  • fault tolerance
  • leader election
  • performance
  • replicated services

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Becker, D., Junqueira, F., & Serafini, M. (2011). Leader election for replicated services using application scores. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 7049 LNCS, pp. 289-308). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7049 LNCS). https://doi.org/10.1007/978-3-642-25821-3_15

Leader election for replicated services using application scores. / Becker, Diogo; Junqueira, Flavio; Serafini, Marco.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7049 LNCS 2011. p. 289-308 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7049 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Becker, D, Junqueira, F & Serafini, M 2011, Leader election for replicated services using application scores. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 7049 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7049 LNCS, pp. 289-308, 12th ACM/IFIP/USENIX International Middleware Conference, Middleware 2011, Lisbon, Portugal, 12/12/11. https://doi.org/10.1007/978-3-642-25821-3_15
Becker D, Junqueira F, Serafini M. Leader election for replicated services using application scores. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7049 LNCS. 2011. p. 289-308. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-25821-3_15
Becker, Diogo ; Junqueira, Flavio ; Serafini, Marco. / Leader election for replicated services using application scores. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 7049 LNCS 2011. pp. 289-308 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{1fa293b508414502abe17d553be15cef,
title = "Leader election for replicated services using application scores",
abstract = "Replicated services often rely on a leader to order client requests and broadcast state updates. In this work, we present POLE, a leader election algorithm that select leaders using application-specific scores. This flexibility given to the application enables the algorithm to tailor leader election according to metrics that are relevant in practical settings and that have been overlooked by existing approaches. Recovery time and request latency are examples of such metrics. To evaluate POLE, we use ZooKeeper, an open-source replicated service used for coordinating Web-scale applications. Our evaluation over realistic wide-area settings shows that application scores can have a significant impact on performance, and that just optimizing the latency of consensus does not translate into lower latency for clients. An important conclusion from our results is that obtaining a general strategy that satisfies a wide range of requirements is difficult, which implies that configurability is indispensable for practical leader election.",
keywords = "fault tolerance, leader election, performance, replicated services",
author = "Diogo Becker and Flavio Junqueira and Marco Serafini",
year = "2011",
month = "12",
day = "23",
doi = "10.1007/978-3-642-25821-3_15",
language = "English",
isbn = "9783642258206",
volume = "7049 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "289--308",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Leader election for replicated services using application scores

AU - Becker, Diogo

AU - Junqueira, Flavio

AU - Serafini, Marco

PY - 2011/12/23

Y1 - 2011/12/23

N2 - Replicated services often rely on a leader to order client requests and broadcast state updates. In this work, we present POLE, a leader election algorithm that select leaders using application-specific scores. This flexibility given to the application enables the algorithm to tailor leader election according to metrics that are relevant in practical settings and that have been overlooked by existing approaches. Recovery time and request latency are examples of such metrics. To evaluate POLE, we use ZooKeeper, an open-source replicated service used for coordinating Web-scale applications. Our evaluation over realistic wide-area settings shows that application scores can have a significant impact on performance, and that just optimizing the latency of consensus does not translate into lower latency for clients. An important conclusion from our results is that obtaining a general strategy that satisfies a wide range of requirements is difficult, which implies that configurability is indispensable for practical leader election.

AB - Replicated services often rely on a leader to order client requests and broadcast state updates. In this work, we present POLE, a leader election algorithm that select leaders using application-specific scores. This flexibility given to the application enables the algorithm to tailor leader election according to metrics that are relevant in practical settings and that have been overlooked by existing approaches. Recovery time and request latency are examples of such metrics. To evaluate POLE, we use ZooKeeper, an open-source replicated service used for coordinating Web-scale applications. Our evaluation over realistic wide-area settings shows that application scores can have a significant impact on performance, and that just optimizing the latency of consensus does not translate into lower latency for clients. An important conclusion from our results is that obtaining a general strategy that satisfies a wide range of requirements is difficult, which implies that configurability is indispensable for practical leader election.

KW - fault tolerance

KW - leader election

KW - performance

KW - replicated services

UR - http://www.scopus.com/inward/record.url?scp=83755188041&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=83755188041&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-25821-3_15

DO - 10.1007/978-3-642-25821-3_15

M3 - Conference contribution

SN - 9783642258206

VL - 7049 LNCS

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

SP - 289

EP - 308

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

ER -