[sunset4] Call for comments regarding NAT64 port allocation

[GangChen]

Wg,

We have some preliminary thoughts for converging those three drafts.
Please kindly check the attached for your reference.
Thank Tom for drafting the slides.
We would like to take the slides as the starting point to prepare the draft.
There are also a couple of questions to the wg in order to guide
authors to write draft.

a) Should we recommend a particular solution in the draft to allocate
nat64 port?
b) Or, should we describe all the known solutions neutrally and
evaluate pros/cons?

Your comments on the slides or above questions are appreciated.

Many thanks

Gang

I guess it's better to open this message to wg and get comments.
Actually, there are several drafts on the table regarding the NAT64
port allocation.
For example, draft-chen-sunset4-cgn-port-allocation,
draft-tsou-behave-natx4-log-reduction and
draft-donley-behave-deterministic-cgn
Submit NAT64 portallocation and address sharing methods to IESG
forconsideration as an Informational RFC
We would like to make a joint presentation in the meeting. If there is
any comments/suggestions for this convergence, we would be much
appreciated.
Best Regards
Gang

Dear chairs,
We intend to the make a joint presentation in Sunset4 for a converged
draft of NAT64-port allocation.
The new draft is not available at this time. We are trying to upload
when there is a clear thought how to combine different drafts.
The presentation in this meeting would like to propose some ideas for
the converged items.
We expect to receive the feedback from wg for the next step.
Best Regards
Gang

[Ted Lemon]

Why not just use PCP?

[GangChen]

I guess PCP could a way to allocate the port. My thought was port
allocation with PCP is the behavior of PCP client+NAT64+PCP server. We
may not expect each NAT64 is capable of PCP server.

BRs

Gang

Post by Ted Lemon
Why not just use PCP?

[Marc Blanchet]

Le 2013-11-04 à 16:36, GangChen <***@gmail.com> a écrit :

cut all cc:. please use the sunset4 ml further.

Post by GangChen
I guess PCP could a way to allocate the port. My thought was port
allocation with PCP is the behavior of PCP client+NAT64+PCP server. We
may not expect each NAT64 is capable of PCP server.

then it is a question of tradeoffs. Needs to be ironed out and discussed.

Marc.

Post by GangChen
BRs
Gang

Post by Ted Lemon
Why not just use PCP?

[GangChen]

Post by Marc Blanchet
cut all cc:. please use the sunset4 ml further.

Post by GangChen
I guess PCP could a way to allocate the port. My thought was port
allocation with PCP is the behavior of PCP client+NAT64+PCP server. We
may not expect each NAT64 is capable of PCP server.

then it is a question of tradeoffs. Needs to be ironed out and discussed.

Yes. It should be discussed.
BTW, there is an analysis draft for PCP, it may provide some
information help to discuss
http://www.ietf.org/id/draft-chen-pcp-mobile-deployment-04.txt

BRs

Gang

Post by Marc Blanchet
Marc.

Post by GangChen
BRs
Gang

Post by Ted Lemon
Why not just use PCP?

[Tom Taylor]

This is about how and why you use PCP.

Post by Ted Lemon
Why not just use PCP?

[Reinaldo Penno (repenno)]

Port or group of ports. There is draft from quite a few folks on the
mechanism to do so.

http://tools.ietf.org/html/draft-ietf-pcp-port-set-03

This draft is in WGLC.

Post by GangChen
I guess PCP could a way to allocate the port. My thought was port
allocation with PCP is the behavior of PCP client+NAT64+PCP server. We
may not expect each NAT64 is capable of PCP server.
BRs
Gang

Post by Ted Lemon
Why not just use PCP?

_______________________________________________
sunset4 mailing list
https://www.ietf.org/mailman/listinfo/sunset4