Narrow Results

One of the interesting features of software systems is that a large number of soft-errors are inherently derated (masked) at the software level. The rate of error-deration may depend on the inherent features of design and programming structures used in the software. This paper investigates the inherent effects of different programming structures on the program resiliency; to this end, six different benchmark programs were implemented by four different programming methods and structures. Profiling experiments were performed on the benchmarks to identify those features of the programming structures that affect the rate of error-deration; then, in order to quantify and examine the inherent resiliency of the programming structures, about 6,720,000 faults were injected into the benchmark programs. The results reveal that about 42.67% of faults in the programs implemented based on while loop-structure are derated; this figure for the programs implemented based on do-while, for and recursive structures are respectively 42.18%, 40.22% and 31.66%. The programs based on the while loop-structure have about 11.01% higher resiliency than the recursive programs. These results can enable software designers to select the most resilient loop-structures for developing inherently resilient programs.

Decreasing the scale of transistors and exponential increase in the transistor counts has made the soft-errors as one of the major causes of software failures. Fault injection is a powerful method for dependability assessment of a computer system against soft-errors. A considerable number of randomly injected faults in the current methods and tools are effect-less or equivalent. To overcome this problem and reduce the cost of fault injection, this study presents a software based fault-injection method that accurately evaluates the dependability of a computer system with a limited number fault-injection. Using a genetic algorithm (GA) the most vulnerable executable paths of an input program is identified; then only the basic blocs (BBs) into the identified vulnerable paths are considered as the target of fault injection. The results of fault injections on the set of 8 traditional benchmark-programs show that the proposed method reduces about 20% of effect-less faults by avoiding the injection of faults in the error-derating blocks of a program. Furthermore, the number of injected faults is reduced to 60% of its original size in the random injection. Also, the proposed method provides more stable and accurate results than the random injection.