Dependability validation, evaluation and testing of safety-critical aerospace systems

Computing systems used in space are required to operate unattended for large periods and survive in the harsh space environment, subject to extremely hostile conditions such as heavy doses of radiation from diverse origins. Radiation, and particularly hits of pro-tons or heavy-ions in space may cause single-event up-sets in electronics, result in transient faults of the computing equipment onboard and compromise its proper function and mission success. Apart from extremely careful design and implementation, safety-critical aerospace systems must incorporate fault tolerance mechanisms in order to recover from faults that inevitably occur in space. The problem of evaluating and validating these mechanisms in the ground in order to qualify them for space missions is becoming increasingly important due to the increasing complexity of both hardware and software used in space missions. The experimental evaluation and validation through fault injection is one approach that has achieved quite success in the past, but some factors impaired traditional fault injection from being used extensively in a broad range of mission architectures. This paper presents, Reception, a technology for evaluating and validating fault tolerance mechanisms, and overall software testing in safety critical systems. Xception injects faults in processor internal functional units, memory, and in general in any device accessible to software. Xception is totally build in software, provides a flexible and powerful set of triggers, is low-intrusive, and may be easily ported to a wide spectrum of systems, thus being an interesting technology for use in safety-critical aerospace systems.