Formal correctness, safety, dependability, and performance analysis of a satellite
This paper reports on the usage of a broad palette of formal modeling and analysis techniques on a regular industrial-size design of an ultra-modern satellite platform. These efforts were carried out in parallel with the conventional software development of the satellite platform. The model itself i...
Saved in:
Published in | 2012 34th International Conference on Software Engineering (ICSE) pp. 1022 - 1031 |
---|---|
Main Authors | , , , , |
Format | Conference Proceeding |
Language | English |
Published |
IEEE
01.06.2012
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | This paper reports on the usage of a broad palette of formal modeling and analysis techniques on a regular industrial-size design of an ultra-modern satellite platform. These efforts were carried out in parallel with the conventional software development of the satellite platform. The model itself is expressed in a formalized dialect of AADL. Its formal nature enables rigorous and automated analysis, for which the recently developed COMPASS toolset was used. The whole effort revealed numerous inconsistencies in the early design documents, and the use of formal analyses provided additional insight on discrete system behavior (comprising nearly 50 million states), on hybrid system behavior involving discrete and continuous variables, and enabled the automated generation of large fault trees (66 nodes) for safety analysis that typically are constructed by hand. The model's size pushed the computational tractability of the algorithms underlying the formal analyses, and revealed bottlenecks for future theoretical research. Additionally, the effort led to newly learned practices from which subsequent formal modeling and analysis efforts shall benefit, especially when they are injected in the conventional software development lifecycle. The case demonstrates the feasibility of fully capturing a system-level design as a single comprehensive formal model and analyze it automatically using a toolset based on (probabilistic) model checkers. |
---|---|
ISBN: | 9781467310666 1467310662 |
ISSN: | 0270-5257 1558-1225 |
DOI: | 10.1109/ICSE.2012.6227118 |