Predicting upgrade failures using dependency analysis

  • Authors:
  • Pietro Abate;Roberto Di Cosmo

  • Affiliations:
  • Universitè Paris Diderot, PPS, UMR 7126, France;Universitè Paris Diderot, PPS, UMR 7126, France

  • Venue:
  • ICDEW '11 Proceedings of the 2011 IEEE 27th International Conference on Data Engineering Workshops
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

Upgrades in component based systems can disrupt other components. Being able to predict the possible consequence of an upgrade just by analysing inter-component dependencies can avoid errors and downtime. In this paper we precisely identify in a repository the components p whose upgrades force a large set of others components to be upgraded. We are also able to discriminate whether all the future versions of p have the same impact, or whether there are different classes of future versions that have different impacts. We perform our analysis on Debian, one of the largest FOSS distributions.