Estimating computational requirements in multi-threaded applications

Juan F. Pérez, Giuliano Casale, Sergio Pacheco-Sanchez

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

Performance models provide effective support for managing quality-of-service (QoS) and costs of enterprise applications. However, expensive high-resolution monitoring would be needed to obtain key model parameters, such as the CPU consumption of individual requests, which are thus more commonly estimated from other measures. However, current estimators are often inaccurate in accounting for scheduling in multi-threaded application servers. To cope with this problem, we propose novel linear regression and maximum likelihood estimators. Our algorithms take as inputs response time and resource queue measurements and return estimates of CPU consumption for individual request types. Results on simulated and real application datasets indicate that our algorithms provide accurate estimates and can scale effectively with the threading levels.

Original languageEnglish (US)
Article number6926798
Pages (from-to)264-278
Number of pages15
JournalIEEE Transactions on Software Engineering
Volume41
Issue number3
DOIs
StatePublished - Mar 1 2015
Externally publishedYes

Fingerprint

Program processors
Linear regression
Maximum likelihood
Quality of service
Servers
Scheduling
Monitoring
Costs
Industry

All Science Journal Classification (ASJC) codes

  • Software

Cite this

Pérez, Juan F. ; Casale, Giuliano ; Pacheco-Sanchez, Sergio. / Estimating computational requirements in multi-threaded applications. In: IEEE Transactions on Software Engineering. 2015 ; Vol. 41, No. 3. pp. 264-278.
@article{a6838ca108de43a59241a7ac65e81b8a,
title = "Estimating computational requirements in multi-threaded applications",
abstract = "Performance models provide effective support for managing quality-of-service (QoS) and costs of enterprise applications. However, expensive high-resolution monitoring would be needed to obtain key model parameters, such as the CPU consumption of individual requests, which are thus more commonly estimated from other measures. However, current estimators are often inaccurate in accounting for scheduling in multi-threaded application servers. To cope with this problem, we propose novel linear regression and maximum likelihood estimators. Our algorithms take as inputs response time and resource queue measurements and return estimates of CPU consumption for individual request types. Results on simulated and real application datasets indicate that our algorithms provide accurate estimates and can scale effectively with the threading levels.",
author = "P{\'e}rez, {Juan F.} and Giuliano Casale and Sergio Pacheco-Sanchez",
year = "2015",
month = "3",
day = "1",
doi = "10.1109/TSE.2014.2363472",
language = "English (US)",
volume = "41",
pages = "264--278",
journal = "IEEE Transactions on Software Engineering",
issn = "0098-5589",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "3",

}

Estimating computational requirements in multi-threaded applications. / Pérez, Juan F.; Casale, Giuliano; Pacheco-Sanchez, Sergio.

In: IEEE Transactions on Software Engineering, Vol. 41, No. 3, 6926798, 01.03.2015, p. 264-278.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Estimating computational requirements in multi-threaded applications

AU - Pérez, Juan F.

AU - Casale, Giuliano

AU - Pacheco-Sanchez, Sergio

PY - 2015/3/1

Y1 - 2015/3/1

N2 - Performance models provide effective support for managing quality-of-service (QoS) and costs of enterprise applications. However, expensive high-resolution monitoring would be needed to obtain key model parameters, such as the CPU consumption of individual requests, which are thus more commonly estimated from other measures. However, current estimators are often inaccurate in accounting for scheduling in multi-threaded application servers. To cope with this problem, we propose novel linear regression and maximum likelihood estimators. Our algorithms take as inputs response time and resource queue measurements and return estimates of CPU consumption for individual request types. Results on simulated and real application datasets indicate that our algorithms provide accurate estimates and can scale effectively with the threading levels.

AB - Performance models provide effective support for managing quality-of-service (QoS) and costs of enterprise applications. However, expensive high-resolution monitoring would be needed to obtain key model parameters, such as the CPU consumption of individual requests, which are thus more commonly estimated from other measures. However, current estimators are often inaccurate in accounting for scheduling in multi-threaded application servers. To cope with this problem, we propose novel linear regression and maximum likelihood estimators. Our algorithms take as inputs response time and resource queue measurements and return estimates of CPU consumption for individual request types. Results on simulated and real application datasets indicate that our algorithms provide accurate estimates and can scale effectively with the threading levels.

UR - http://www.scopus.com/inward/record.url?scp=84925135842&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84925135842&partnerID=8YFLogxK

U2 - 10.1109/TSE.2014.2363472

DO - 10.1109/TSE.2014.2363472

M3 - Article

AN - SCOPUS:84925135842

VL - 41

SP - 264

EP - 278

JO - IEEE Transactions on Software Engineering

JF - IEEE Transactions on Software Engineering

SN - 0098-5589

IS - 3

M1 - 6926798

ER -