Static Analysis of SEU Effects on Software Applications

Control Flow Errors have been widely addressed in literature as a possible threat to the dependability of computer systems, and many clever techniques have been proposed to detect and tolerate them. Nevertheless, it has never been discussed i f the overheads introduced by many of these techniques ore justified by a reasonable probabilip of incurring Control Flow Errors. This paper presents a static executable code analysis methodology able to compute, depending on the torget microprocessor platform, the upper-boundprobability that o given application incurs in a Control Flow Error