Adaptive peer sampling with newscast

Norbert Tölgyesi, M. Jelasity

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

27 Citations (Scopus)

Abstract

The peer sampling service is a middleware service that provides random samples from a large decentralized network to support gossip-based applications such as multicast, data aggregation and overlay topology management. Lightweight gossip-based implementations of the peer sampling service have been shown to provide good quality random sampling while also being extremely robust to many failure scenarios, including node churn and catastrophic failure. We identify two problems with these approaches. The first problem is related to message drop failures: if a node experiences a higher-than-average message drop rate then the probability of sampling this node in the network will decrease. The second problem is that the application layer at different nodes might request random samples at very different rates which can result in very poor random sampling especially at nodes with high request rates. We propose solutions for both problems. We focus on Newscast, a robust implementation of the peer sampling service. Our solution is based on simple extensions of the protocol and an adaptive self-control mechanism for its parameters, namely-without involving failure detectors-nodes passively monitor local protocol events using them as feedback for a local control loop for self-tuning the protocol parameters. The proposed solution is evaluated by simulation experiments.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages523-534
Number of pages12
Volume5704 LNCS
DOIs
Publication statusPublished - 2009
EventEuro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings - Delft, Netherlands
Duration: Aug 25 2009Aug 28 2009

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5704 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

OtherEuro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings
CountryNetherlands
CityDelft
Period8/25/098/28/09

Fingerprint

Sampling
Vertex of a graph
Gossip
Random Sampling
Network protocols
Failure Detectors
Data Aggregation
Self-tuning
Middleware
Multicast
Overlay
Decentralized
Simulation Experiment
Monitor
Agglomeration
Tuning
Topology
Detectors
Feedback
Decrease

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Tölgyesi, N., & Jelasity, M. (2009). Adaptive peer sampling with newscast. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5704 LNCS, pp. 523-534). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5704 LNCS). https://doi.org/10.1007/978-3-642-03869-3_50

Adaptive peer sampling with newscast. / Tölgyesi, Norbert; Jelasity, M.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5704 LNCS 2009. p. 523-534 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5704 LNCS).

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

Tölgyesi, N & Jelasity, M 2009, Adaptive peer sampling with newscast. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 5704 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 5704 LNCS, pp. 523-534, Euro-Par 2009 Parallel Processing - 15th International Euro-Par Conference, Proceedings, Delft, Netherlands, 8/25/09. https://doi.org/10.1007/978-3-642-03869-3_50
Tölgyesi N, Jelasity M. Adaptive peer sampling with newscast. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5704 LNCS. 2009. p. 523-534. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-03869-3_50
Tölgyesi, Norbert ; Jelasity, M. / Adaptive peer sampling with newscast. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 5704 LNCS 2009. pp. 523-534 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{2f19c38954ec4a7f859add8c8a751c76,
title = "Adaptive peer sampling with newscast",
abstract = "The peer sampling service is a middleware service that provides random samples from a large decentralized network to support gossip-based applications such as multicast, data aggregation and overlay topology management. Lightweight gossip-based implementations of the peer sampling service have been shown to provide good quality random sampling while also being extremely robust to many failure scenarios, including node churn and catastrophic failure. We identify two problems with these approaches. The first problem is related to message drop failures: if a node experiences a higher-than-average message drop rate then the probability of sampling this node in the network will decrease. The second problem is that the application layer at different nodes might request random samples at very different rates which can result in very poor random sampling especially at nodes with high request rates. We propose solutions for both problems. We focus on Newscast, a robust implementation of the peer sampling service. Our solution is based on simple extensions of the protocol and an adaptive self-control mechanism for its parameters, namely-without involving failure detectors-nodes passively monitor local protocol events using them as feedback for a local control loop for self-tuning the protocol parameters. The proposed solution is evaluated by simulation experiments.",
author = "Norbert T{\"o}lgyesi and M. Jelasity",
year = "2009",
doi = "10.1007/978-3-642-03869-3_50",
language = "English",
isbn = "3642038689",
volume = "5704 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "523--534",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Adaptive peer sampling with newscast

AU - Tölgyesi, Norbert

AU - Jelasity, M.

PY - 2009

Y1 - 2009

N2 - The peer sampling service is a middleware service that provides random samples from a large decentralized network to support gossip-based applications such as multicast, data aggregation and overlay topology management. Lightweight gossip-based implementations of the peer sampling service have been shown to provide good quality random sampling while also being extremely robust to many failure scenarios, including node churn and catastrophic failure. We identify two problems with these approaches. The first problem is related to message drop failures: if a node experiences a higher-than-average message drop rate then the probability of sampling this node in the network will decrease. The second problem is that the application layer at different nodes might request random samples at very different rates which can result in very poor random sampling especially at nodes with high request rates. We propose solutions for both problems. We focus on Newscast, a robust implementation of the peer sampling service. Our solution is based on simple extensions of the protocol and an adaptive self-control mechanism for its parameters, namely-without involving failure detectors-nodes passively monitor local protocol events using them as feedback for a local control loop for self-tuning the protocol parameters. The proposed solution is evaluated by simulation experiments.

AB - The peer sampling service is a middleware service that provides random samples from a large decentralized network to support gossip-based applications such as multicast, data aggregation and overlay topology management. Lightweight gossip-based implementations of the peer sampling service have been shown to provide good quality random sampling while also being extremely robust to many failure scenarios, including node churn and catastrophic failure. We identify two problems with these approaches. The first problem is related to message drop failures: if a node experiences a higher-than-average message drop rate then the probability of sampling this node in the network will decrease. The second problem is that the application layer at different nodes might request random samples at very different rates which can result in very poor random sampling especially at nodes with high request rates. We propose solutions for both problems. We focus on Newscast, a robust implementation of the peer sampling service. Our solution is based on simple extensions of the protocol and an adaptive self-control mechanism for its parameters, namely-without involving failure detectors-nodes passively monitor local protocol events using them as feedback for a local control loop for self-tuning the protocol parameters. The proposed solution is evaluated by simulation experiments.

UR - http://www.scopus.com/inward/record.url?scp=70350676811&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70350676811&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-03869-3_50

DO - 10.1007/978-3-642-03869-3_50

M3 - Conference contribution

AN - SCOPUS:70350676811

SN - 3642038689

SN - 9783642038686

VL - 5704 LNCS

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

SP - 523

EP - 534

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

ER -