Re: [v6ops] Implementation Status of PREF64

[Vasilenko Eduard <[e-mail]>]

Ø  I don’t know of anyone whose desire for IPv6 PAT has delayed enterprise adoption beyond the point where the ability to use stateful inspection without NAT was adequately explained to them.
Then how to deal with the MHMP environment?
If the host has 2 addresses from different providers (for redundancy) then how to properly choose combination {source IP, next hop} for a particular session?
Any implementation of RFC 8028?
IPv4 was using Private IP for such site with ad-hoc NAT translation on gateways to specific Carrier.

Ed/
From: v6ops [mailto:[e-mail]] On Behalf Of Owen DeLong
Sent: Saturday, October 9, 2021 7:08 PM
To: David Farmer <farmer=[e-mail]>
Cc: v6ops list <[e-mail]>; Owen DeLong <owen=[e-mail]>
Subject: Re: [v6ops] Implementation Status of PREF64




On Oct 8, 2021, at 14:32 , David Farmer <farmer=[e-mail]<mailto:farmer=[e-mail]>> wrote:



On Fri, Oct 8, 2021 at 3:36 PM Owen DeLong <owen=[e-mail]<mailto:[e-mail]>> wrote:



On Oct 8, 2021, at 9:28 AM, David Farmer <farmer=[e-mail]<mailto:farmer=[e-mail]>> wrote:

I will note a few things;

1. The IETF hasn't picked a value of  N, but it has, through RFC7934, said that N<1.

I doubt that. N<1 would be utterly useless… Perhaps N>1?

Yes, I meant N>1, but let's explore N<1 for a moment ;

One could argue that means NAT, and more specifically Port-Translation NAT.

Ewww



So, does that mean the IETF needs to specify and equipment vendors implement Port-Translation NAT for IPv6? I know a lot of network operators, especially enterprise networks, that at least say they want that. And, you are saying the network operators should have ultimate and unilateral control.

Please no. We didn’t specify it (until it was widely deployed) in IPv4 and I think sending the signal that it SHOULD NOT be standardized in IPv6 is perfectly fine.

In IPv4, it was a work around for an address shortage. If IPv6 experiences such an address shortage, I’ll accept that we might need to consider such a thing. Short of that, code is available, operators are free to deploy it if they really want it, but I don’t think we should sanction it.

OTOH, IA_NA is already sanctioned and android’s failure to implement it is delaying IPv6 adoption in the enterprise.

I don’t know of anyone whose desire for IPv6 PAT has delayed enterprise adoption beyond the point where the ability to use stateful inspection without NAT was adequately explained to them.


Again, network operators, public and private, have, and should have, a lot of control, but it is not ultimate and unilateral control.

Fine, but IA_NA is standardized and SHOULD be implemented on any IPv6 host with sufficient capabilities to do so.

Owen