On Distribution and Impact of Fault Effects at Real-Time Kernel and Application Levels

Embedded systems as well as real-time kernels for their construction are sensitive to transient and other faults, each of which can lead to various errors at various system levels and can potentially result in the system failure. In existing works, sensitivity to transient SEU errors was studied only with the goal to classify SEU impacts to the system behavior. In this paper, the study is extended to permanent faults being injected into the persistent memory used to store the program and into the volatile memory used to store data structures. Moreover, it is studied there how the sensitivity is affected if combination of selected software-implemented fault tolerant techniques is built into the kernel and task levels of the system. The set of common fault tolerant techniques is extended to a novel watchdog-inspired technique designed to check the correctness of task executions. On the experimental basis, it is shown that the fault impacts can be softened by the proper combination of the techniques.