摘要：I point to some challenges for WCET analysis offered in the transition to integrated mixed-criticality systems (MCSs) and to multi-core platforms, claim that proposed certification standards are inadequate, show that the MCS model heavily used by the scheduling community is fraught, and clarify why the traditional abstract interface between WCET analysis and schedulability analysis is obsolete. A central point is the insistence on sound approaches. I give a detailed account of how the most rigid certification procedures, those of the avionics domain, are satisfied, to defend the validity of my claims.
