|
H. Hecht, M. Hecht, "Qualitative Interpretation of Software Test Data", Computer-Aided
Design, Test and Evaluation for Dependability Workshop, Beijing, China, July, 1996
During the final test phase of a high integrity software project the incidence of
failures may be so low that conventional reliability growth models become unusable
because the stochastics of the failure process mask any decreasing failure trend.
While the low failure rate is of course desirable, it is frequently not sufficient
by itself to demonstrate compliance with reliability or availability requirements.
Particularly for software used in safety systems, the maximum failure rate may be
specified as low as 10-10 per hour, a value that cannot be verified in practice
by conventional demonstration tests. We identify a qualitative approach to test
data interpretation, particularly the examination of rare conditions as causes for
failures, as a possible avenue for reliability assessment. This can be used as an
alternative or as a supplement to redundancy for achieving the highest possible
level of reliability for high integrity software. Further research in this area
is recommended.
Click to Download
|
|
