[v6ops] Re: New Version Notification for draft-link-v6ops-claton-03.txt

[Tommy Jensen <[e-mail]>]

Hello Masanobu-¤µ¤ó,

Jen jumped on this ¡ª you may find the text here: https://github.com/furry13/v6ops-464xlat-enable/pull/27/files

You will note we agreed on MUST rather than SHOULD ¡ª I feel rather strongly about this as it does not affect the request for a configuration knob either way but highlights the risks of doing so by default. That said, I'm open to discussion on that point, and if we end up at a stand-off, what others in the WG think.

Thanks,
Tommy

________________________________
From: Kawashima Masanobu(´¨u¡¡ÕýÉì)
Sent: Sunday, June 16, 2024 10:57 PM
To: Tommy Jensen
Cc: Jen Linkova; [e-mail]; V6 Ops List
Subject: [EXTERNAL] Re: New Version Notification for draft-link-v6ops-claton-03.txt




Hi Tommy,

Thank you for your reply.


>Then we are on the same page. My point was that end users are

£¾unlikely to go turning on CLAT at scale, that it will be the exception.

Just FYI.
In Japan, BIGLOBE is providing NAT64 service for wireline.
It has approximately 300 thousand of subscribers.
The amount is more than a few.
There is a motivation to turn on CLAT, not using native IPv4(from PPPoE).
Therefore, I prefer SHOULD rather than MUST.

>Jen and I will need to add some consideration text to accompany it

>to explain why this escape hatch is not recommended general practice,

>but that is a solvable problem.

I¡¯d like to review the text. :-)


Regards,

Masanobu



========================

 NEC Platforms, Ltd.

 KAWASHIMA Masanobu

 [e-mail]<mailto:[e-mail]>

 https://www.necplatforms.co.jp/en/

========================







From: Tommy Jensen <Jensen.Thomas=[e-mail]>
Sent: Thursday, June 13, 2024 11:38 PM
To: Kawashima Masanobu(´¨u¡¡ÕýÉì) <[e-mail]>
Cc: Jen Linkova <[e-mail]>; [e-mail]; V6 Ops List <[e-mail]>
Subject: Re: [EXTERNAL] [v6ops] Re: ##freemail## Re: Re: New Version Notification for draft-link-v6ops-claton-03.txt



(trimming long email)



Hello Masanobu-¤µ¤ó,



>If CE router is purchased at retail shop, ISP can¡¯t control it.
>We¡¯re not sure if the CE router support Option 108.



Ah, this was the part I was missing. The CE is not in the ISP's control.



>If not, admin(ISP or end user) is going to enable CLAT somehow.
>This is the exceptional case.



Then we are on the same page. My point was that end users are

unlikely to go turning on CLAT at scale, that it will be the exception.

If you are ok with that, that platforms are not going to have this as a

default configuration, then I think I am ok with the suggested change.



Jen and I will need to add some consideration text to accompany it

to explain why this escape hatch is not recommended general practice,

but that is a solvable problem.



Thanks,

Tommy



________________________________

From: Kawashima Masanobu(´¨u¡¡ÕýÉì)
Sent: Thursday, June 13, 2024 2:05 AM
To: Tommy Jensen
Cc: Jen Linkova; [e-mail]<mailto:[e-mail]>; V6 Ops List
Subject: RE: [EXTERNAL] [v6ops] Re: ##freemail## Re: Re: New Version Notification for draft-link-v6ops-claton-03.txt





Hi Tommy,

Thank you for your reply.
I¡¯m not sure if I understand your question correctly.

My point is that hosts don¡¯t know the upstream is IPv4 PPPoE
or IPv4aaS or both. CE router only knows it.

For example, when CE router provide IPv4 private address to hosts
and upstream is IPv4 PPPoE, then also provive IPv6 GUA to hosts
and upstream is IPv4aaS such as 464XLAT at the same time.

In this situation, admin(typically ISP) prefer let hosts use CLAT.
Because IPv4aaS is faster than IPv4 PPPoE in Japan.

If CE router is purchased at retail shop, ISP can¡¯t control it.
We¡¯re not sure if the CE router support Option 108.
If not, admin(ISP or end user) is going to enable CLAT somehow.
This is the exceptional case.

>> Maybe we can say "MUST disable CLAT unless explicitly configured not to do so"?

>> Would that address your case?

So, the sentence is reasonable or SHOULD is also fine.

Regards,

Masanobu



========================

 NEC Platforms, Ltd.

 KAWASHIMA Masanobu

 [e-mail]<mailto:[e-mail]>

 https://www.necplatforms.co.jp/en/

========================