This white paper introduces datapath analysis, a method of design analysis for the data processing elements used in functional safety and other mission critical markets that have been designed without the specific requirements of a functional safety system, which is also known as safety element out of context. This method is presented as an alternative to the industry-accepted failure mode effects analysis (FMEA) method, with a specific focus on components or elements designed out-of-context for general applications with an emphasis on functional safety.
Datapath analysis is a refinement of FMEA and is used to proactively identify where probable faults might occur and justify where diagnostics for supporting functional safety applications are required for an intellectual property (IP) design. Unlike FMEA, datapath analysis leverages the four principles of data processing that map potential failure modes to potential diagnostics for elements designed and built without context. Unlike datapath analysis, FMEA requires a severity index and an occurrence probability of cause so the failure modes can be prioritized. Because these components are designed out of context, the severity of one failure over another is not known because any failure might be catastrophic. The paramount functional safety metric driving quality is intended operation. Any assumptions regarding severity are restrictive because knowing every use case for context might not be possible.