Excerpts from Christian Vogt on Wed, Feb 25, 2009 07:16:15PM -0800:
> On Feb 21, 2009, Scott Brim wrote:
>
>>> 3.1.2.  Translation
>>>
>>>   Translation solutions are characterized by a translation
>>>   operation between an identifier to a locator and back to an
>>>   identifier as the packet traverses the network.
>>
>> Now that I've pondered this a little, I can't think of any approach
>> that does that (translate from identifier to locator and back again).
>
>
> Scott -
>
> All address indirection solutions that are application-transparent do
> map both source and destination addresses at both ends.  I.e., you have
> four mappings per packets:  The respective local addresses (source
> address at the sending side; destination address at the destination
> side) are mapped to achieve provider independence.  The respective
> remote addresses are mapped to achieve application transparency.

You just helped make my point.  The sentence says translation is from
an "identifier" to a "locator" and back again.  The scenario you
describe is not that, it's between global and local addresses.

>> NAT-based approaches and Six/One Router are also one-way: the local
>> source address is translated to an RLOC when outgoing, and the
>> destination address is translated to a local address when incoming.
>
> As you mention Six/One Router -- this solution proposal actually has two
> modes:  One mode is NOT application-transparent; this, as you say, maps
> only the respective local addresses (two mappings per packet).  But
> Six/One Router also has an application-transparent mode, which performs
> all four mappings per packet.  The two modes are called Unilateral mode
> and Bilateral mode, respectively.
>
> And FWIW:  Six/One Router, Unilateral mode == NAT66.

Agreed.
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to