Operating system function reuse to achieve low-cost fault tolerance

The aim of this article is to propose a new approach to fault tolerance in single processor embedded systems which is centred on the operating system. Particular attention is put on low-cost techniques that exploit functions already present in the system in a different than-usual way to achieve protection with little or no intervention at the application level. An example is given: the realization of a checkpoint and rollback scheme through the context switch function.