DocumentCode :
358714
Title :
Investigating software failures with a software black box
Author :
Elbaum, Sebastian ; Munson, John C.
Author_Institution :
Dept. of Comput. Sci. & Eng., Nebraska Univ., Lincoln, NE, USA
Volume :
4
fYear :
2000
fDate :
2000
Firstpage :
547
Abstract :
One of the greatest safety improvement inventions for the airline industry has been the crash-protected Flight Data Recorder (FDR). Today, the FDR are mandatory equipment in most civil aircraft. With the data retrieved from the FDR, the last moments before an accident can be reconstructed. Constructing the analog of the FDR for avionics software would be very beneficial. When complex systems fail, it is often very difficult to determine the precise cause of the failure. This is largely due to insufficient or inappropriate data collection process, which does not permit the reconstruction of the circumstances that led to the failure. This research effort presents the Software Black Box (SBB), which constitutes a framework that facilitates the investigation and understanding of software failures. The SBB specifies a mechanism to capture the essentials of an executing program, and it provides a reconstruction technique that allows the generation of the scenarios that may have led to the software failure. The SBB architecture and a validation study are presented. The results provide insights into the tradeoffs and potential of the SBB
Keywords :
aerospace computing; avionics; data structures; safety-critical software; supervisory programs; avionics software; behavioral model; black box decoder; call graph; data structures; executing program; reconstruction technique; scenarios generation; software black box; software failures; transition frequency matrix; transition probability matrix; Accidents; Aerospace electronics; Aircraft; Computer crashes; Computer industry; Information retrieval; Pervasive computing; Protection; Real time systems; Software systems;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Aerospace Conference Proceedings, 2000 IEEE
Conference_Location :
Big Sky, MT
ISSN :
1095-323X
Print_ISBN :
0-7803-5846-5
Type :
conf
DOI :
10.1109/AERO.2000.878756
Filename :
878756
Link To Document :
بازگشت