Detail publikace

Noise Injection Heuristics for Concurrency Testing

Originální název

Noise Injection Heuristics for Concurrency Testing

Anglický název

Noise Injection Heuristics for Concurrency Testing

Jazyk

en

Originální abstrakt

Testing of concurrent software is difficult due tothe non-determinism present in scheduling of concurrent threads. Existingtesting approaches tackle this problem either using a modified scheduler whichallows to systematically explore possible scheduling alternatives or usingrandom or heuristic noise injection which allows to observe different schedulingscenarios. In this paper, we experimentally compare several existing noiseinjection heuristics both from the point of view of coverage of possiblebehaviours as well as from the point of view of error discovery probability.Moreover, we also propose a new noise injection heuristics which usesconcurrency coverage information to decide where to put noise and show that itcan outperform the existing approaches in certain cases.

Anglický abstrakt

Testing of concurrent software is difficult due tothe non-determinism present in scheduling of concurrent threads. Existingtesting approaches tackle this problem either using a modified scheduler whichallows to systematically explore possible scheduling alternatives or usingrandom or heuristic noise injection which allows to observe different schedulingscenarios. In this paper, we experimentally compare several existing noiseinjection heuristics both from the point of view of coverage of possiblebehaviours as well as from the point of view of error discovery probability.Moreover, we also propose a new noise injection heuristics which usesconcurrency coverage information to decide where to put noise and show that itcan outperform the existing approaches in certain cases.

BibTex


@article{BUT91448,
  author="Bohuslav {Křena} and Zdeněk {Letko} and Tomáš {Vojnar}",
  title="Noise Injection Heuristics for Concurrency Testing",
  annote="Testing of concurrent software is difficult due tothe non-determinism present in
scheduling of concurrent threads. Existingtesting approaches tackle this problem
either using a modified scheduler whichallows to systematically explore possible
scheduling alternatives or usingrandom or heuristic noise injection which allows
to observe different schedulingscenarios. In this paper, we experimentally
compare several existing noiseinjection heuristics both from the point of view of
coverage of possiblebehaviours as well as from the point of view of error
discovery probability.Moreover, we also propose a new noise injection heuristics
which usesconcurrency coverage information to decide where to put noise and show
that itcan outperform the existing approaches in certain cases.",
  address="NEUVEDEN",
  chapter="91448",
  edition="NEUVEDEN",
  howpublished="print",
  institution="NEUVEDEN",
  number="7119",
  volume="2012",
  year="2012",
  month="january",
  pages="123--131",
  publisher="NEUVEDEN",
  type="journal article - other"
}