Automatic performance model generation for java enterprise edition (EE) applications

Andreas Brunnert, Christian Vögele, Helmut Krcmar

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

28 Scopus citations

Abstract

The effort required to create performance models for enterprise applications is often out of proportion compared to their benefits. This work aims to reduce this effort by introducing an approach to automatically generate component-based performance models for running Java EE applications. The approach is applicable for all Java EE server products as it relies on standardized component types and interfaces to gather the required data for modeling an application. The feasibility of the approach and the accuracy of the generated performance models are evaluated in a case study using a SPECjEnterprise2010 industry standard benchmark deployment. Simulations based on a generated performance model of this reference deployment show a prediction error of 1 to 20 % for response time and of less than 10 % for CPU utilization and throughput.

Original languageEnglish
Title of host publicationComputer Performance Engineering - 10th European Workshop, EPEW 2013, Proceedings
PublisherSpringer Verlag
Pages74-88
Number of pages15
ISBN (Print)9783642407246
DOIs
StatePublished - 2013
Event10th European Workshop on Performance Engineering, EPEW 2013 - Venice, Italy
Duration: 16 Sep 201317 Sep 2013

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume8168 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th European Workshop on Performance Engineering, EPEW 2013
Country/TerritoryItaly
CityVenice
Period16/09/1317/09/13

Keywords

  • Enterprise Applications
  • Java Enterprise Edition
  • Palladio Component Model
  • Performance Evaluation
  • Performance Modeling

Fingerprint

Dive into the research topics of 'Automatic performance model generation for java enterprise edition (EE) applications'. Together they form a unique fingerprint.

Cite this