Software Approach to Hazard Detection Using On-line Analysis of Safety Constraints

The research here addresses the problem of improving software safety through hazard detection. The premise of our work is that hazard situations can and do occur, and are often complex, involving multiple sources. So there is a need for a mechanism to detect complex hazards and react in a timely and meaningful way. This paper addresses such a detection mechanism through Cnet, an on-line analysis tool that supports the specification of complex multi-source hazards using a query-like language, uses both synchronous and asynchronous checking approaches to balance efficiency and expressiveness, accommodates dynamic applications through dynamic constraint addition, and supports distributed and parallel applications running in heterogeneous environments