[address-policy-wg] Mandating NAT toward the final /8

Sascha Lenz wrote:

> waering all my hats as private internet citizen, network operator for my own network, network architect for a non-commerial organisation supporting new technologies, LIR, ISP, and IT Consultant for big to large customers, i hearby state that i ...

To be a good internet citizen, you should read Saltzer's paper on
end to end argument.

> a) have operational IPv6 networks for 6 years now

How many end users do you have?

> b) don't like NAT

I don't like legacy NAT too because it is not complete nor correct.

A direct consequence of the end to end argument is:

   NAT can completely and correctly be implemented only with the
   knowledge and help of the end hosts behind a NAT gateway.

which is not the case with legacy NAT.

> f) don't see a reason to conserve IPv4 space
> g) don't see a reason not to migrate to IPv6

The problem, here, is that there is no path to migrate to IPv6.

Or, do you still believe dual stack approach work?

> i) don't support any more complex NAT setup than we already have in the 
> wild now

Fully agreed. Keep It Simple, Stupid.

How do you think about proposals from people desperately working
to deploy IPv6 to have complex and stateful NAT between 4 and 6?

							Masataka Ohta