Gossip-based distributed algorithms for estimating the average load of scalable clusters and grids

Amnon Barak*, Zvi Drezner

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

5 Scopus citations

Abstract

We present gossip-based distributed algorithms by which each node of a cluster or a grid can find an estimate of the global average load of the system. Knowledge of this estimate can improve the performance by allowing each node to make better scheduling decisions. Our algorithms are based on asynchronous exchanges (gossip) of load messages between either a random or a fixed pairs of nodes. We show that based on these messages, each node can find an estimate of the average load of all the nodes. We present the algorithms and prove their convergence.

Original languageEnglish
Title of host publicationProceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications, PDPTA'04
EditorsH.R. Arabnia, J. Ni
Pages610-616
Number of pages7
StatePublished - 2004
EventProceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications, PDPTA'04 - Las Vegas, NV, United States
Duration: 21 Jun 200424 Jun 2004

Publication series

NameProceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications, PDPTA'04
Volume2

Conference

ConferenceProceedings of the International Conference on Parallel and Distributed Processing Techniques and Applications, PDPTA'04
Country/TerritoryUnited States
CityLas Vegas, NV
Period21/06/0424/06/04

Keywords

  • Cluster computing
  • Distributed computing
  • Gossip-based algorithms
  • Grid computing

Fingerprint

Dive into the research topics of 'Gossip-based distributed algorithms for estimating the average load of scalable clusters and grids'. Together they form a unique fingerprint.

Cite this