Got pinged by Alain on draft-penno-softwire-sdnat status...

[Francis Dupont]

> Alain dropped me an email yesterday regarding the new
> draft-penno-softwire-sdnat-02.txt spec.
> 
> He was essentially asking "which version were we supporting"

=> we have first to understand where are the differences (I have
a diff window on them (http://tools.ietf.org/rfcdiff)).

> I think the appropriate internal "marketing answer" is:
> 
> 1) for IETF in Paris "we are doing that we are doing now. We are just
> focused on getting the demo working"
> 
> 2) for post-IETF code, "we will move to version 2"

=> it is more more to version 3 because the ICMP stuff in the spec
is deeply broken.

> Since it is you doing the actual coding, what should I actually tell him ?

=> we support the idea of the spec (the SD-CPE is port restricted,
the SD-CGN is stateless) in the DS-Lite context (so it is SD-B4
and SD-AFTR) but not the (re)config via ICMP, nor compatibility modes
(the "stateless" is global, not per tunnel/B4).

Regards

Francis Dupont <fdupont at isc.org>

PS: my opinion about the ICMP stuff is my own opinion:
 - I am not in charge to code it
 - anyway the smooth reconfig of a SD-B4 is a hard problem which
  won't be solved in phase 1
 - we still have to propose something better for the config
  (BTW the problem is there already are too many proposals :-).
If it is not clear, the config we'll use for the demo will be a static one.
PPS: if Alain really insists to get the ICMP stuff, ask for an official
type value assigned by IANA (:-)...