On Tue, 25 Nov 2008, Keith Moore wrote:
Iljitsch van Beijnum wrote:
On 23 nov 2008, at 20:25, Tony Hain wrote:
The fundamental problem here is that the voices of those bearing the
costs
in the core are being represented, while the voices of those doing
application development are not being heard.
Not sure that's entirely true...
But in any event, compared to the backflips through flaming hoops we
have to do in IPv4, the asking a remote server what our source address
looks like from the outside to make address based referrals work doesn't
seem too onerous. Or do you disagree?
absolutely it's too onerous. why in the world should an application's
deployability depend on the existence of a server that lives in global
address space -- or for that matter, on a bank of servers that exist to
do nothing but forward traffic? isn't that what the network is supposed
to do?
Actually, he did not say the server forwarded traffic, just that it
provided a way to learn how 'my' address was mapped through 'my' nat. A
ip-ip mapping lookup service, probably not all that different than DNS or
the service the telcos use to map 8xx numbers to actual phones, or phone
numbers to specific terminal devices on specific carriers.
I don't know if that is a good approach or not but I find it quite a bit
less onerous than routing all traffic via an intermediate server. And it
seemed the question wasn't understood.
Dave Morris
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf