2012/2/2 Rémi Després <[email protected]> > > > 2. (because) RFC6145 has provided (not thorough but) good enough > transparency when used in double translation, and accordingly it is NOT > needed to be updated; > > > Full transparency, based on clear theory, remains AFAIK better than "good > enough" based on some collected statistic at a given time. >
as i stated, we want to have better transparency but only if it is not costly and the solution has no other obvious flaws. > > 3. carrying ICMPv4 directly in IPv6 payload will be a harmful idea and > less feasible. > > > I don't see where you see that ICMPv4 payload tunneled in a mapped header > isn't easily feasible. > not feasible in term of supporting the operation where the single/double translations are needed simultaneously and are needed to be treated statelessly. > > Stating it is harmful is of course not a proof. > Discussion on this point can continue with more details in other e-amils > based on configurations you have in mind. > i have proposed a concrete problem rather than a simple statement at least twice in the past mail dialogues. ;-) - maoke
_______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
