Consistent detection of global predicates under a weak fault assumption

We study the problem of detecting general global predicates in distributed systems where all application processes and at most t<m monitor processes may be subject to crash faults, where m is the total number of monitor processes in the system. We introduce two new observation modalities called negotiably and discernibly (which correspond to possibly and definitely in fault-free systems) and present detection algorithms for them which work under increasingly weak fault assumptions