On Nov 30, 2012, at 6:37 PM, Randy Bush wrote: >> Actually, Randy, engineering with the upper bound is pretty typical. > > from your document > > "The calculations in this document, while just candidate representations > of RPKI, project a very long lower-bound on the potential time needed > for RP caches to gather all objects in the RPKI"
lol... Good point: computing the upper bound on time, using the lower bound on size... Different subjects, both conservative estimates, but my bad for being unclear. :) >> The counter tends to be viewed as quite amateurish. > > personally, i would be professionally embarrassed to have my name on > such wild assed cabbage throwing as that document. That's very constructive, thank you! >> codifying a system's design after building an implementation, and then >> retrofitting requirements to that design so that we can _then_ back >> into a threat model that befits a system's capabilities is typically >> frowned upon by engineers. > > let's be honest here. the threat model was done in the late '80s. the > base of the design in the early '90s. we're just slogging through the > saussage machine. Um... Kind of a lot has changed since then, and I _believe_ the design was rejected back then too... I think it might be less professionally embarrassing to ask, what needs to change to make this viable? ;) ymmv, Eric _______________________________________________ sidr mailing list [email protected] https://www.ietf.org/mailman/listinfo/sidr
