Le 2012-04-03 à 11:31, Satoru Matsushima a écrit :

> On 2012/04/03, at 18:07, Rémi Després wrote:
> 
>> 
>> Le 2012-04-03 à 03:53, Satoru Matsushima a écrit :
>> 
>>> FYI, choosing MAP doesn't mean that committing to a 'single'. But choosing 
>>> 4rd-u means that committing to a 'single'.
>>> 
>>> There're just transport variants, which are encapsulation, translation and 
>>> new one we've never seen before. Proponents of the new one claim to unify 
>>> transport only to the new one, which means that eliminating both 
>>> encapsulation and translation variants of MAP because they say two choice 
>>> of transport type would make confusion for operators.
>> 
>>> The new one is transparent than translation, but less than encapsulation.
>> 
>> If you would have a detailed scenario where ANY transparency difference 
>> could be noticeable e2e between a MAP-E tunnel and a 4rd-U tunnel, this 
>> sentence would be fair.
>> However, you haven't shown such a scenario, and AFAIK you can't have one 
>> because it doesn't exist.
>> The above assertion, being wrong, is therefore misleading.
>> 
> 
> I already noticed rfc5082 as a detail.

AFAIK, rfc5082 compatibility is ensured with TTL preservation that is in 
4rd-u-06.
Anything more?

> 
> 
>> 
>>> I couldn't decide transport type to be unified with the new one.
>> 
>>> No evidence, no expertise, no experience for that.
>> 
>> But many of those who have no motivation to look for flaws that don't exist, 
>> do understand that the design is safe.
>> 
> 
> I can't understand what you mean. Let us study it in your BOF.

There is no such thing as "my BOF"!


>>> Do we need them to be sure in spite of we already have existing mature 
>>> transport variants?
>> 
>> MAP-E+T isn't as mature as repeatedly claimed.
>> MAP-T may be more imprecise than MAP-E in this respect, but both have known 
>> bugs to be fixed (at least packet IDs of shared-addres CEs of MAP-T, and BR 
>> hairpinning in MAP-E).
>> 
> 
> thanks, the authors of map suite are working to improve the documents.

> Don't worry about encapsulation and translation.

Just being tired to hear that the spec if clean and complete, and to have to 
find out myself what is missing.


>> Besides, the exact list of MAP DHCPv6 parameters has yet to be finalized.

Thanks for having implicitly acknowledged this fact (contrary to the claim that 
MAP is completely specified).

> 
> Besides, 4rd-u is on the stage of pre-BOF discussion.

Sorry to read this kind of argument from you.
No other comment.

RD



> 
> cheers,
> --satoru

_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires

Reply via email to