Load sharing based on task resource prediction

Kumar K. Goswami, Ravishankar K. Iyer, Murthy Devarakonda

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

2 Citations (Scopus)

Abstract

Predicted task resource usage provides a basis for developing two centralized load-sharing policies: MinQ and MinResp. Trace-driven simulations are used to compare MinQ and MinResp against Centex, an efficient centralized scheme. Experimental results show that the use of prediction makes MinQ and MinResp significantly less sensitive to the status update rate than Centex. Consequently, the proposed algorithms perform better than Centex at slower update rates and are capable of handling larger workloads. The prediction-based policies are also highly effective for load-sharing in environments with widely varying CPU requirements. Using a real trace file, with an equal number of large and small tasks, MinResp consistently produced mean response times that were 9% to 35% lower than those of Centex.

Original languageEnglish (US)
Title of host publicationProceedings of the Hawaii International Conference on System Science
EditorsBruce D. Shriver
PublisherPubl by Western Periodicals Co
Pages921-927
Number of pages7
Volume2
ISBN (Print)0818619120
StatePublished - Dec 1 1989
Externally publishedYes
EventProceedings of the Twenty-Second Annual Hawaii International Conference on System Sciences: Software Track - Kailua-Kona, Hawaii, USA
Duration: Jan 3 1989Jan 6 1989

Other

OtherProceedings of the Twenty-Second Annual Hawaii International Conference on System Sciences: Software Track
CityKailua-Kona, Hawaii, USA
Period1/3/891/6/89

Fingerprint

Program processors

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Goswami, K. K., Iyer, R. K., & Devarakonda, M. (1989). Load sharing based on task resource prediction. In B. D. Shriver (Ed.), Proceedings of the Hawaii International Conference on System Science (Vol. 2, pp. 921-927). Publ by Western Periodicals Co.

Load sharing based on task resource prediction. / Goswami, Kumar K.; Iyer, Ravishankar K.; Devarakonda, Murthy.

Proceedings of the Hawaii International Conference on System Science. ed. / Bruce D. Shriver. Vol. 2 Publ by Western Periodicals Co, 1989. p. 921-927.

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

Goswami, KK, Iyer, RK & Devarakonda, M 1989, Load sharing based on task resource prediction. in BD Shriver (ed.), Proceedings of the Hawaii International Conference on System Science. vol. 2, Publ by Western Periodicals Co, pp. 921-927, Proceedings of the Twenty-Second Annual Hawaii International Conference on System Sciences: Software Track, Kailua-Kona, Hawaii, USA, 1/3/89.
Goswami KK, Iyer RK, Devarakonda M. Load sharing based on task resource prediction. In Shriver BD, editor, Proceedings of the Hawaii International Conference on System Science. Vol. 2. Publ by Western Periodicals Co. 1989. p. 921-927
Goswami, Kumar K. ; Iyer, Ravishankar K. ; Devarakonda, Murthy. / Load sharing based on task resource prediction. Proceedings of the Hawaii International Conference on System Science. editor / Bruce D. Shriver. Vol. 2 Publ by Western Periodicals Co, 1989. pp. 921-927
@inproceedings{dd70e5b8b0d74a17bdd8c3168d509439,
title = "Load sharing based on task resource prediction",
abstract = "Predicted task resource usage provides a basis for developing two centralized load-sharing policies: MinQ and MinResp. Trace-driven simulations are used to compare MinQ and MinResp against Centex, an efficient centralized scheme. Experimental results show that the use of prediction makes MinQ and MinResp significantly less sensitive to the status update rate than Centex. Consequently, the proposed algorithms perform better than Centex at slower update rates and are capable of handling larger workloads. The prediction-based policies are also highly effective for load-sharing in environments with widely varying CPU requirements. Using a real trace file, with an equal number of large and small tasks, MinResp consistently produced mean response times that were 9{\%} to 35{\%} lower than those of Centex.",
author = "Goswami, {Kumar K.} and Iyer, {Ravishankar K.} and Murthy Devarakonda",
year = "1989",
month = "12",
day = "1",
language = "English (US)",
isbn = "0818619120",
volume = "2",
pages = "921--927",
editor = "Shriver, {Bruce D.}",
booktitle = "Proceedings of the Hawaii International Conference on System Science",
publisher = "Publ by Western Periodicals Co",

}

TY - GEN

T1 - Load sharing based on task resource prediction

AU - Goswami, Kumar K.

AU - Iyer, Ravishankar K.

AU - Devarakonda, Murthy

PY - 1989/12/1

Y1 - 1989/12/1

N2 - Predicted task resource usage provides a basis for developing two centralized load-sharing policies: MinQ and MinResp. Trace-driven simulations are used to compare MinQ and MinResp against Centex, an efficient centralized scheme. Experimental results show that the use of prediction makes MinQ and MinResp significantly less sensitive to the status update rate than Centex. Consequently, the proposed algorithms perform better than Centex at slower update rates and are capable of handling larger workloads. The prediction-based policies are also highly effective for load-sharing in environments with widely varying CPU requirements. Using a real trace file, with an equal number of large and small tasks, MinResp consistently produced mean response times that were 9% to 35% lower than those of Centex.

AB - Predicted task resource usage provides a basis for developing two centralized load-sharing policies: MinQ and MinResp. Trace-driven simulations are used to compare MinQ and MinResp against Centex, an efficient centralized scheme. Experimental results show that the use of prediction makes MinQ and MinResp significantly less sensitive to the status update rate than Centex. Consequently, the proposed algorithms perform better than Centex at slower update rates and are capable of handling larger workloads. The prediction-based policies are also highly effective for load-sharing in environments with widely varying CPU requirements. Using a real trace file, with an equal number of large and small tasks, MinResp consistently produced mean response times that were 9% to 35% lower than those of Centex.

UR - http://www.scopus.com/inward/record.url?scp=0024860879&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0024860879&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0024860879

SN - 0818619120

VL - 2

SP - 921

EP - 927

BT - Proceedings of the Hawaii International Conference on System Science

A2 - Shriver, Bruce D.

PB - Publ by Western Periodicals Co

ER -