Compositional Generation of Software Architecture Performance QN Models

  • Authors:
  • Affiliations:
  • Venue:
  • WICSA '04 Proceedings of the Fourth Working IEEE/IFIP Conference on Software Architecture
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Early performance analysis based on Queueing NetworkModels (QNM) has been often proposed to support softwaredesigners during the software development process. Theseapproaches aim at addressing performance issues as earlyas possible in order to reduce design failures. All of themtry to adapt to software systems the well-known system performanceanalysis methodology. This implies that they assumeat design time the availability of information about thehardware platform the software will run on. In recent yearswe have proposed a methodology that allows quantitativereasoning on software aspects without considering hardwareaspects. In this work we extend our methodology toencompass a compositional approach to performance analysisof software architecture described by means of UML2.0diagrams. The main improvements include the characterizationof architectural patterns and of their correspondingQNM pattern; the use of multi-chain queueing network assystem target model and the identification of the informationneeded to parameterize the system model.