Real-time scheduling of divisible loads in clusters to handle estimated execution time inaccuracies

Due Luong, Ying Lu, Jitender Deogun, Steve Godclard

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

Abstract

Quality of Service (QoS) provisioning for divisible loads in clusters, enabled using real-time scheduling theory, is based on an important assumption: the scheduler knows the execution time of even' task in the workload. Information from production clusters, however, shows that user estimated task execution times are often inaccurate. In this paper, we present an approach to address this problem, where a feedback mechanism is developed to overcome the inaccuracies in user estimation of task execution time. Besides QoS provisioning of divisible loads in clusters, the proposed algorithm achieves low task reject ratio and high system utilization. Simulation evaluation shows that the new algorithm outperforms the previous approach.

Original languageEnglish (US)
Title of host publicationIMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings
Pages242-247
Number of pages6
StatePublished - Dec 1 2010
Event3rd International Multi-Conference on Engineering and Technological Innovation, IMETI 2010 - Orlando, FL, United States
Duration: Jun 29 2010Jul 2 2010

Publication series

NameIMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings
Volume2

Conference

Conference3rd International Multi-Conference on Engineering and Technological Innovation, IMETI 2010
CountryUnited States
CityOrlando, FL
Period6/29/107/2/10

Fingerprint

Quality of service
Scheduling
Feedback
Provisioning
Feedback mechanism
Workload
Simulation
Evaluation

ASJC Scopus subject areas

  • Management of Technology and Innovation
  • Engineering(all)

Cite this

Luong, D., Lu, Y., Deogun, J., & Godclard, S. (2010). Real-time scheduling of divisible loads in clusters to handle estimated execution time inaccuracies. In IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings (pp. 242-247). (IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings; Vol. 2).

Real-time scheduling of divisible loads in clusters to handle estimated execution time inaccuracies. / Luong, Due; Lu, Ying; Deogun, Jitender; Godclard, Steve.

IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings. 2010. p. 242-247 (IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings; Vol. 2).

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

Luong, D, Lu, Y, Deogun, J & Godclard, S 2010, Real-time scheduling of divisible loads in clusters to handle estimated execution time inaccuracies. in IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings. IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings, vol. 2, pp. 242-247, 3rd International Multi-Conference on Engineering and Technological Innovation, IMETI 2010, Orlando, FL, United States, 6/29/10.
Luong D, Lu Y, Deogun J, Godclard S. Real-time scheduling of divisible loads in clusters to handle estimated execution time inaccuracies. In IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings. 2010. p. 242-247. (IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings).
Luong, Due ; Lu, Ying ; Deogun, Jitender ; Godclard, Steve. / Real-time scheduling of divisible loads in clusters to handle estimated execution time inaccuracies. IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings. 2010. pp. 242-247 (IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings).
@inproceedings{75b5446ad62e43edb82dfe755ced42d5,
title = "Real-time scheduling of divisible loads in clusters to handle estimated execution time inaccuracies",
abstract = "Quality of Service (QoS) provisioning for divisible loads in clusters, enabled using real-time scheduling theory, is based on an important assumption: the scheduler knows the execution time of even' task in the workload. Information from production clusters, however, shows that user estimated task execution times are often inaccurate. In this paper, we present an approach to address this problem, where a feedback mechanism is developed to overcome the inaccuracies in user estimation of task execution time. Besides QoS provisioning of divisible loads in clusters, the proposed algorithm achieves low task reject ratio and high system utilization. Simulation evaluation shows that the new algorithm outperforms the previous approach.",
author = "Due Luong and Ying Lu and Jitender Deogun and Steve Godclard",
year = "2010",
month = "12",
day = "1",
language = "English (US)",
isbn = "9781936338030",
series = "IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings",
pages = "242--247",
booktitle = "IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings",

}

TY - GEN

T1 - Real-time scheduling of divisible loads in clusters to handle estimated execution time inaccuracies

AU - Luong, Due

AU - Lu, Ying

AU - Deogun, Jitender

AU - Godclard, Steve

PY - 2010/12/1

Y1 - 2010/12/1

N2 - Quality of Service (QoS) provisioning for divisible loads in clusters, enabled using real-time scheduling theory, is based on an important assumption: the scheduler knows the execution time of even' task in the workload. Information from production clusters, however, shows that user estimated task execution times are often inaccurate. In this paper, we present an approach to address this problem, where a feedback mechanism is developed to overcome the inaccuracies in user estimation of task execution time. Besides QoS provisioning of divisible loads in clusters, the proposed algorithm achieves low task reject ratio and high system utilization. Simulation evaluation shows that the new algorithm outperforms the previous approach.

AB - Quality of Service (QoS) provisioning for divisible loads in clusters, enabled using real-time scheduling theory, is based on an important assumption: the scheduler knows the execution time of even' task in the workload. Information from production clusters, however, shows that user estimated task execution times are often inaccurate. In this paper, we present an approach to address this problem, where a feedback mechanism is developed to overcome the inaccuracies in user estimation of task execution time. Besides QoS provisioning of divisible loads in clusters, the proposed algorithm achieves low task reject ratio and high system utilization. Simulation evaluation shows that the new algorithm outperforms the previous approach.

UR - http://www.scopus.com/inward/record.url?scp=84890839720&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84890839720&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84890839720

SN - 9781936338030

T3 - IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings

SP - 242

EP - 247

BT - IMETI 2010 - 3rd International Multi-Conference on Engineering and Technological Innovation, Proceedings

ER -