DocumentCode
2209391
Title
Reliability estimation of hierarchical software systems
Author
Leblanc, S.P. ; Roman, P.A.
Author_Institution
R. Mil. Coll. of Canada, Kingston, Ont., Canada
fYear
2002
fDate
2002
Firstpage
249
Lastpage
253
Abstract
The reliability evaluation of hardware systems is usually well integrated into the design process. Because it is done early, this reliability evaluation is useful in making design decisions. Software Reliability Evaluation (SRE), on the other hand, has been mostly conducted after development has been completed, therefore offering little or no input into the design. This paper proposes a simple approach to estimate the reliability of software systems that are composed of a hierarchy of modules. Furthermore, the proposed approach takes advantage of the familiar reliability block diagram (RBD) methodology, which has long been used for hardware reliability estimation. This paper demonstrates how RBD can be used to represent hierarchical software systems. Using this technique early in the development process can offer system designers the opportunity to conduct sensitivity analysis that combines both hardware and software components. Criticality is also shown to be useful in the proposed SRE approach. Examples are given to show how criticality makes the SRE more realistic than the use of RBD alone
Keywords
hierarchical systems; program testing; software reliability; criticality; hierarchical software systems reliability estimation; module hierarchy; reliability block diagram methodology; software components; software development process; Educational institutions; Hardware; Object oriented modeling; Process design; Programming; Sensitivity analysis; Software reliability; Software systems; Statistical distributions; Vocabulary;
fLanguage
English
Publisher
ieee
Conference_Titel
Reliability and Maintainability Symposium, 2002. Proceedings. Annual
Conference_Location
Seattle, WA
ISSN
0149-144X
Print_ISBN
0-7803-7348-0
Type
conf
DOI
10.1109/RAMS.2002.981650
Filename
981650
Link To Document