[arin-ppml] The role of NAT in IPv6

[Chris Engel]

Olivier MJ Crépin-Leblond wrote:

> I'd be really grateful if this matter could be put to bed. I
> mean - it's obvious that there are two sides which don't
> agree. This discussion can go on ad-infinitum and I somehow
> don't see a solution to this. I'm usually not someone to say
> this, and I humbly hope that I shan't offend anyone, but in
> this instance, it's becoming repetitive...

Oliver,

You are correct in the above. More to the point, it's rather academic to the purposes of this list. The needs of some-one who wants private IPv6 space for the purposes of using it in conjunction with NAT are not (I believe) functionaly different from those who want private IPv6 for other purposes. As long as people are in general agreement that IPv6 private space should be available for people who want it (for whatever purposes) then whether those organizations choose to deploy NAT or not....seems entirely tangential.

With ULA-random that's entirely a non-issue (as it seems use-able in roughly the equivalent fashion that RFC1918 space currently is).

The only issue that I could see potentialy cropping up would be if there was some particular conflict in the justification rules for ULA-C. Not sure exactly what the details of those would be.... but essentialy getting ULA-C assignment for ones equipment shouldn't generaly hinder/count against getting an assignment of Public IP address space for a subset of that same equipment. That's about the only issue I could see that would be relevant to the current discussion.


Scott Leibrand wrote:

> Chris,
>
> Can you describe the 1:many PAT use case for IPv6? I get the
> need for
> 1:1 NAT, but not for 1:many PAT.

Scott,

I'll reply off-list in order to spare folks the gory details of this discussion. As it seems like this discussion is causing uneccesary clutter for many folks here. Anyone that's actualy interested in this issue...feel free to drop me a note and I'll CC you on my answer to Scott.





Christopher Engel