Competitive analysis of task scheduling algorithms on a fault-prone machine and the impact of resource augmentation

Antonio Fernández Anta, Chryssis Georgiou, Dariusz R. Kowalski, Elli Zavou

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

Abstract

Reliable task execution on machines that are prone to unpredictable crashes and restarts is both important and challenging, but not much work exists on the analysis of such systems. We consider the online version of the problem, with tasks arriving over time at a single machine under worst-case assumptions. We analyze the fault-tolerant properties of four popular scheduling algorithms: Longest In System (LIS), Shortest In System (SIS), Largest Processing Time (LPT) and Shortest Processing Time (SPT). We use three metrics for the evaluation and comparison of their competitive performance, namely, completed load, pending load, and latency. We also investigate the effect of resource augmentation in their performance, by increasing the speed of the machine. Hence, we compare the behavior of the algorithms for different speed intervals and show that there is no clear winner with respect to all the three considered metrics. While SPT is the only algorithm that achieves competitiveness on completed load for small speed, LIS is the only one that achieves competitiveness on latency (for large enough speed).

Original languageEnglish (US)
Title of host publicationAdaptive Resource Management and Scheduling for Cloud Computing - 2nd International Workshop, ARMS-CC 2015 Held in Conjunction with ACM Symposium on Principles of Distributed Computing, PODC 2015 Donostia-San Sebastián, Spain, July 20, 2015, Revised Selected Papers
EditorsFlorin Pop, Maria Potop-Butucaru
PublisherSpringer Verlag
Pages1-16
Number of pages16
ISBN (Print)9783319284477
DOIs
StatePublished - Jan 1 2015
Externally publishedYes
Event2nd International Workshop on Adaptive Resource Management and Scheduling for Cloud Computing, ARMS-CC 2015 - Donostia-San Sebastián, Spain
Duration: Jul 20 2015Jul 20 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9438
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference2nd International Workshop on Adaptive Resource Management and Scheduling for Cloud Computing, ARMS-CC 2015
CountrySpain
CityDonostia-San Sebastián
Period7/20/157/20/15

Fingerprint

Resource Augmentation
Competitive Analysis
Task Scheduling
Scheduling algorithms
Scheduling Algorithm
Fault
Competitiveness
Latency
Processing
Metric
Restart
Single Machine
Crash
Fault-tolerant
Interval
Evaluation

Keywords

  • Competitive analysis
  • Failures
  • Online algorithms
  • Resource augmentation
  • Scheduling
  • Task sizes

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Fernández Anta, A., Georgiou, C., Kowalski, D. R., & Zavou, E. (2015). Competitive analysis of task scheduling algorithms on a fault-prone machine and the impact of resource augmentation. In F. Pop, & M. Potop-Butucaru (Eds.), Adaptive Resource Management and Scheduling for Cloud Computing - 2nd International Workshop, ARMS-CC 2015 Held in Conjunction with ACM Symposium on Principles of Distributed Computing, PODC 2015 Donostia-San Sebastián, Spain, July 20, 2015, Revised Selected Papers (pp. 1-16). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9438). Springer Verlag. https://doi.org/10.1007/978-3-319-28448-4_1

Competitive analysis of task scheduling algorithms on a fault-prone machine and the impact of resource augmentation. / Fernández Anta, Antonio; Georgiou, Chryssis; Kowalski, Dariusz R.; Zavou, Elli.

Adaptive Resource Management and Scheduling for Cloud Computing - 2nd International Workshop, ARMS-CC 2015 Held in Conjunction with ACM Symposium on Principles of Distributed Computing, PODC 2015 Donostia-San Sebastián, Spain, July 20, 2015, Revised Selected Papers. ed. / Florin Pop; Maria Potop-Butucaru. Springer Verlag, 2015. p. 1-16 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9438).

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

Fernández Anta, A, Georgiou, C, Kowalski, DR & Zavou, E 2015, Competitive analysis of task scheduling algorithms on a fault-prone machine and the impact of resource augmentation. in F Pop & M Potop-Butucaru (eds), Adaptive Resource Management and Scheduling for Cloud Computing - 2nd International Workshop, ARMS-CC 2015 Held in Conjunction with ACM Symposium on Principles of Distributed Computing, PODC 2015 Donostia-San Sebastián, Spain, July 20, 2015, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9438, Springer Verlag, pp. 1-16, 2nd International Workshop on Adaptive Resource Management and Scheduling for Cloud Computing, ARMS-CC 2015, Donostia-San Sebastián, Spain, 7/20/15. https://doi.org/10.1007/978-3-319-28448-4_1
Fernández Anta A, Georgiou C, Kowalski DR, Zavou E. Competitive analysis of task scheduling algorithms on a fault-prone machine and the impact of resource augmentation. In Pop F, Potop-Butucaru M, editors, Adaptive Resource Management and Scheduling for Cloud Computing - 2nd International Workshop, ARMS-CC 2015 Held in Conjunction with ACM Symposium on Principles of Distributed Computing, PODC 2015 Donostia-San Sebastián, Spain, July 20, 2015, Revised Selected Papers. Springer Verlag. 2015. p. 1-16. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-28448-4_1
Fernández Anta, Antonio ; Georgiou, Chryssis ; Kowalski, Dariusz R. ; Zavou, Elli. / Competitive analysis of task scheduling algorithms on a fault-prone machine and the impact of resource augmentation. Adaptive Resource Management and Scheduling for Cloud Computing - 2nd International Workshop, ARMS-CC 2015 Held in Conjunction with ACM Symposium on Principles of Distributed Computing, PODC 2015 Donostia-San Sebastián, Spain, July 20, 2015, Revised Selected Papers. editor / Florin Pop ; Maria Potop-Butucaru. Springer Verlag, 2015. pp. 1-16 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{babde0f73154441c88b6626506423482,
title = "Competitive analysis of task scheduling algorithms on a fault-prone machine and the impact of resource augmentation",
abstract = "Reliable task execution on machines that are prone to unpredictable crashes and restarts is both important and challenging, but not much work exists on the analysis of such systems. We consider the online version of the problem, with tasks arriving over time at a single machine under worst-case assumptions. We analyze the fault-tolerant properties of four popular scheduling algorithms: Longest In System (LIS), Shortest In System (SIS), Largest Processing Time (LPT) and Shortest Processing Time (SPT). We use three metrics for the evaluation and comparison of their competitive performance, namely, completed load, pending load, and latency. We also investigate the effect of resource augmentation in their performance, by increasing the speed of the machine. Hence, we compare the behavior of the algorithms for different speed intervals and show that there is no clear winner with respect to all the three considered metrics. While SPT is the only algorithm that achieves competitiveness on completed load for small speed, LIS is the only one that achieves competitiveness on latency (for large enough speed).",
keywords = "Competitive analysis, Failures, Online algorithms, Resource augmentation, Scheduling, Task sizes",
author = "{Fern{\'a}ndez Anta}, Antonio and Chryssis Georgiou and Kowalski, {Dariusz R.} and Elli Zavou",
year = "2015",
month = "1",
day = "1",
doi = "10.1007/978-3-319-28448-4_1",
language = "English (US)",
isbn = "9783319284477",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "1--16",
editor = "Florin Pop and Maria Potop-Butucaru",
booktitle = "Adaptive Resource Management and Scheduling for Cloud Computing - 2nd International Workshop, ARMS-CC 2015 Held in Conjunction with ACM Symposium on Principles of Distributed Computing, PODC 2015 Donostia-San Sebasti{\'a}n, Spain, July 20, 2015, Revised Selected Papers",

}

TY - GEN

T1 - Competitive analysis of task scheduling algorithms on a fault-prone machine and the impact of resource augmentation

AU - Fernández Anta, Antonio

AU - Georgiou, Chryssis

AU - Kowalski, Dariusz R.

AU - Zavou, Elli

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Reliable task execution on machines that are prone to unpredictable crashes and restarts is both important and challenging, but not much work exists on the analysis of such systems. We consider the online version of the problem, with tasks arriving over time at a single machine under worst-case assumptions. We analyze the fault-tolerant properties of four popular scheduling algorithms: Longest In System (LIS), Shortest In System (SIS), Largest Processing Time (LPT) and Shortest Processing Time (SPT). We use three metrics for the evaluation and comparison of their competitive performance, namely, completed load, pending load, and latency. We also investigate the effect of resource augmentation in their performance, by increasing the speed of the machine. Hence, we compare the behavior of the algorithms for different speed intervals and show that there is no clear winner with respect to all the three considered metrics. While SPT is the only algorithm that achieves competitiveness on completed load for small speed, LIS is the only one that achieves competitiveness on latency (for large enough speed).

AB - Reliable task execution on machines that are prone to unpredictable crashes and restarts is both important and challenging, but not much work exists on the analysis of such systems. We consider the online version of the problem, with tasks arriving over time at a single machine under worst-case assumptions. We analyze the fault-tolerant properties of four popular scheduling algorithms: Longest In System (LIS), Shortest In System (SIS), Largest Processing Time (LPT) and Shortest Processing Time (SPT). We use three metrics for the evaluation and comparison of their competitive performance, namely, completed load, pending load, and latency. We also investigate the effect of resource augmentation in their performance, by increasing the speed of the machine. Hence, we compare the behavior of the algorithms for different speed intervals and show that there is no clear winner with respect to all the three considered metrics. While SPT is the only algorithm that achieves competitiveness on completed load for small speed, LIS is the only one that achieves competitiveness on latency (for large enough speed).

KW - Competitive analysis

KW - Failures

KW - Online algorithms

KW - Resource augmentation

KW - Scheduling

KW - Task sizes

UR - http://www.scopus.com/inward/record.url?scp=84955259625&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84955259625&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-28448-4_1

DO - 10.1007/978-3-319-28448-4_1

M3 - Conference contribution

AN - SCOPUS:84955259625

SN - 9783319284477

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

SP - 1

EP - 16

BT - Adaptive Resource Management and Scheduling for Cloud Computing - 2nd International Workshop, ARMS-CC 2015 Held in Conjunction with ACM Symposium on Principles of Distributed Computing, PODC 2015 Donostia-San Sebastián, Spain, July 20, 2015, Revised Selected Papers

A2 - Pop, Florin

A2 - Potop-Butucaru, Maria

PB - Springer Verlag

ER -