Change-point detection for black-box services

  • Authors:
  • Ilenia Epifani;Carlo Ghezzi;Giordano Tamburrelli

  • Affiliations:
  • Politecnico di Milano, Milano, Italy;Politecnico di Milano, Milano, Italy;Politecnico di Milano, Milano, Italy

  • Venue:
  • Proceedings of the eighteenth ACM SIGSOFT international symposium on Foundations of software engineering
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Modern software systems are increasingly built out of services that are developed, deployed, and operated by independent organizations, which expose them for use by potential clients. Services may be directly invoked by clients. They may also be composed by service integrators, who in turn expose the composite artifact as a new service. Continuous change is typical of this world. Providers may change services and the deployment infrastructure to meet continuously changing requirements and be more competitive. Clients may change their operational profiles. Changes have a severe impact on the quality of services. In this paper we address the problem of identifying changes concerning the non-functional behavior of software services managed by external organizations, and consequently considered as black-box artifacts. We define the concept of change-point and provide a statistical technique aimed at identifying it, given an execution trace produced by client invocations. Change-point detection is key to reasoning about changes, diagnosing their cause, and suitably reacting to their occurrence.