ABSTRACT
Current methods for Technology Qualification (TQ) rely to a large degree on traditional reliability methods developed for hardware systems. However, as technology becomes more dependent upon software, leading to more complex systems, there is a need to better reflect the system perspective. One aspect of complexity is that the system constituents is heavily interconnected and dependent upon each other, i.e. the system shows emergent behavior. Systemic failures, not caused by component failures, but caused by how the constituents interact (i.e. emergence), need to be addressed differently than previous practices. This paper presents a consistent way of classifying failures to help discovering the critical failures and to guide how to handle the identified failures in a system TQ setting. To ensure that the identification process is consistent and as complete as possible, we advocate that failures should be classified according to different perspectives, where the classification within each perspective is Mutually Exclusive and Collectively Exhaustive (MECE). We show how this can be used to guide how evidence is collected in the technology qualification process.
