Achieving Representative Faultloads in Software Fault Injection

摘要: Given the complexity of modern software systems and its pervasiveness in many aspects our lives, faults (i.e., bugs) are a dangerous threat. Unfortunately, it is impossible to assure that perfect despite advances engineering. Therefore, mission- safety-critical have provide fault tolerance algorithms mechanisms mitigate this Software Fault Injection emerged last decades as means for testing improving fault-tolerant systems. This approach deliberately introduces order assess behavior presence faults. In be adopted by practitioners development critical systems, an effective trustworthy evaluation tolerance, realism being injected (fault representativeness) need assured, i.e., should reflect residual escape process can affect system. thesis addresses representativeness with respect three aspects. First, proposes selecting code locations which inject complex The identifies more likely hide from testing, focus injection on most representative reduce number cost experiments at same time. Second, method accuracy binary code, required when source not available case third-party software. Finally, technique emulating concurrency faults, significant part affecting These contributions instrumental advance make practical developing