Use this URL to cite or link to this record in EThOS:
Title: Delayed failure of software components using stochastic testing
Author: Gardiner, J.
ISNI:       0000 0004 2717 9178
Awarding Body: Cranfield University
Current Institution: Cranfield University
Date of Award: 2012
Availability of Full Text:
Access from EThOS:
Access from Institution:
The present research investigates the delayed failure of software components and addresses the problem that the conventional approach to software testing is unlikely to reveal this type of failure. Delayed failure is defined as a failure that occurs some time after the condition that causes the failure, and is a consequence of long-latency error propagation. This research seeks to close a perceived gap between academic research into software testing and industrial software testing practice by showing that stochastic testing can reveal delayed failure, and supporting this conclusion by a model of error propagation and failure that has been validated by experiment. The focus of the present research is on software components described by a request-response model. Within this conceptual framework, a Markov chain model of error propagation and failure is used to derive the expected delayed failure behaviour of software components. Results from an experimental study of delayed failure of DBMS software components MySQL and Oracle XE using stochastic testing with random generation of SQL are consistent with expected behaviour based on the Markov chain model. Metrics for failure delay and reliability are shown to depend on the characteristics of the chosen experimental profile. SQL mutation is used to generate negative as well as positive test profiles. There appear to be few systematic studies of delayed failure in the software engineering literature, and no studies of stochastic testing related to delayed failure of software components, or specifically to delayed failure of DBMS. Stochastic testing is shown to be an effective technique for revealing delayed failure of software components, as well as a suitable technique for reliability and robustness testing of software components. These results provide a deeper insight into the testing technique and should lead to further research. Stochastic testing could provide a dependability benchmark for component-based software engineering.
Supervisor: Fletcher, G. Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available