simFI: From single to simultaneous software fault injections

Abstract

Abstract-Software-implemented fault injection (SWIFI) is an established experimental technique to evaluate the robustness of software systems. While a large number of SWIFI frameworks exist, virtually all are based on a single-fault assumption, i.e., interactions of simultaneously occurring independent faults are not investigated. As software systems containing more than a single fault often are the norm than an exception [1] and current safety standards require the consideration of "multi-point faults"[2], the validity of this single-fault assumption is at question for contemporary software systems.To address the issue and support simultaneous SWIFI (simFI), we analyze how independent faults can manifest in a generic software composition model and extend an existing SWIFI tool to support some characteristic simultaneous fault types. We implement three simultaneous fault models and demonstrate their utility in evaluating the robustness of the Windows CE kernel. Our findings indicate that simultaneous fault injections prove highly efficient in triggering robustness vulnerabilities.

Simultaneous Fault Models for the Generation of Efficient Error Detection Mechanisms

“…M n }, and CON = {(M a k , M a l )}, where M k exports to the import interface of M l over connection a. The adopted system model is consistent with [10] and compatible with the simultaneous fault models developed in [20].…”

“…The simultaneous fault models evaluated in this paper were proposed in [20] in response to a proliferation of software fault injection frameworks making the single-fault assumption, despite this being known to be unrealistic [8]. The models were developed on notions of coincidence and impact before being evaluated with regard to utility using metrics such as coverage and failure induction.…”

“…The simultaneous fault models described were developed to improve software fault injection by overcoming the single-fault assumption, thus permitting more meaningful analyses [20]. As the fault injection conducted in [20] focused on the point of entry to modules, the fault injection experiments in this paper focus on generating EDMs at the entry points to modules, an approach supported by existing research [26].…”

“…As the fault injection conducted in [20] focused on the point of entry to modules, the fault injection experiments in this paper focus on generating EDMs at the entry points to modules, an approach supported by existing research [26]. To maintain compatibility with existing research, software state was characterised by all variables in scope at the point of fault injection.…”

N-version Software Module Requirements to Grant the Software Execution Fault-Tolerance

“… N-version software program modules have been developed taking into account the proposed requirements;  The NVX input/output components do not interact with any external objects and have been designed to emulate such interaction;  Data processed by N-version software have been modelled , for the purposes of testing and definition of the NVX properties;  Errors occurring in the program modules have been injected at the development stage and designed for testing and analysis of the consequences for the NVX and N-version software [22], [23].…”