On 11/09/2012 08:21, Fred Baker (fred) wrote:
> On Sep 11, 2012, at 12:06 AM, Brian E Carpenter wrote:
>
>> o An application instance wishing to establish communication with a
>> peer "behind" an NPTv6 Translator may need to use a different
>> address to reach that peer depending on whether the instance is
>> behind the same NPTv6 Translator or external to it.
>
> a problem Happy Eyeballs easily addresses, and which RFC 3484 and 3484-bis do
> not.
>
>> my point is that encouraging NPTv6 propagates this set
>> of problems in the IPv6 world. They aren't going to go away
>> along with IPv4.
>
> Well, yes, but since edge networks are *not* getting PA addresses from each
> of their upstream providers and putting a /64 on each LAN from each such
> address, but instead getting PI address space, we have no scalable solution
> for that issue.
>
> What's that rule called? TAANSTAAFL, as I recall. Pretending won't make it
> otherwise. Either the service providers deal with the complexities of the
> IPv4 route table, or the edge networks deal with the complexities of managing
> prefixes from each of multiple upstream PA prefixes, or DNS advertises inside
> and outside addresses somehow. Pick one.
This being the Internet, I expect we'll pick several. But indeed, we have
to deal with current reality and also try to influence long term trends
in a desirable way. I think that getting the problem area well described
is the first step.
Brian
_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area
