I don't have much to add to the current discussion that I haven't already said in previous discussion, but do have one additional datum:
A few months ago I said that it seemed possible that the validation-reconsidered approach would alleviate some potential race conditions involved in the loosely consistent database model. For what it's worth, I would like to report that in the one timing-related corner case we managed to explore all the way down to the end of the rat hole, it turned out that validation-reconsidered was irrelevant, because the normal certificate revocation process during a resource reduction produced the same effect regardless of the order in which the RP saw the RPKI objects in question. In other words: no smoking gun found. _______________________________________________ sidr mailing list [email protected] https://www.ietf.org/mailman/listinfo/sidr
