A fault-tolerant program is one that can continue service after faults have adversely affected the state of the program. The process by which a program can tolerate faults is divided into four phases: error detection, when the fault is first detected; damage assessment, when the damage caused by the fault is assessed; fault removal, when the fault is diagnosed, then removed; finally, error recovery, when the impact of the fault on the state of the program is corrected. In this article, we are interested in the first two phases. We discuss means to carry these phases out, using executable assertions. Appropriate executable assertions are derived by solving algebraic equations in the Tarski calculus of relations.
All Science Journal Classification (ASJC) codes
- Information Systems
- Hardware and Architecture