Re: [3gv6] [BEHAVE] DNS64 Resolvers and Dual-Stack Hosts, draft-wing-behave-dns64-config-00

[David Crowe <[e-mail]>]

This is really an absurd statement. Nobody can require old protocols 
to predict the future. They can try, they can do some protection, but 
they'll usually get it wrong in some significant ways.

It's the responsibility of new protocols to deal with the known past 
not the responsibility of old protocols to deal with an unknown 
future.

- David Crowe



At 9:49 AM +1300 1/13/10, Brian E Carpenter wrote:
>Dan,
>
>On 2010-01-12 15:52, Dan Wing wrote:
>> 
>>
>>>  -----Original Message-----
>>>  From: Behcet Sarikaya [mailto:[e-mail]]
>
>...
>>>  You may be right. I personally am not a fan of solutions that
>>>  are not mobility friendly also requiring modifications to
>>>  well established services like DNS.
>>
>>  Yes, IPv6/IPv4 translation has side-effects.  Wish it wasn't
>>  true, but IPv6 was not designed to be upgradable from IPv4.
>>  If it was, we wouldn't be in the current predicament.
>
>Just to be clear, the defect there is in IPv4. IPv4 was not
>designed to be upgradable (e.g. by allowing for variable
>length addresses). So we are condemned by history to choose
>between dual stack and packet translation. Translation is
>for cases where dual stack cannot be applied, and as Dan says
>it has side effects. We've known that for 15 years.
>
>>
>>>  What happened to the solutions developed in the years of
>>>  2000, when the transition was a hot issue?
>>
>>  Are you referring to NAT-PT (RFC2766), published in February
>>  2000?  RFC4966 happened to it, "Reasons to Move the Network
>>  Address Translator - Protocol Translator (NAT-PT) to Historic
>>  Status", http://tools.ietf.org/html/rfc4966.  Several months
>>  after that, BEHAVE had its charter extended to cover IPv6/
>>  IPv4 translation.  NAT-PT also munged with DNS, in a worse
>>  way than DNS64.
>>
>>>  No offense to you or your co-authors, though.
>>
>>  No offense taken.  These discussions are important.
>
>What we are doing now is trying to minimise the side effects,
>building on experience since NAT-PT was designed.
>It's worth reading draft-penno-behave-64-analysis to
>understand which side effects we can't elminate.
>
>     Brian