On Sep 22, 2011, at 1:48 AM, Daniel Park wrote:
> Quick question at this stage, several years back, v6ops formed several
> internal-groups to develop IPv6 transition scenario, and the unmanaged
> network (it meant home network) was published as RFC 3904. Are there any
> relationship between this draft and that RFC efforts or reference points ?
There is some good background in RFC 3904, but it is admittedly incomplete (see
4.1.3, 4.3.4...), goes a bit beyond our scope in some areas, and not far enough
in others.
Breaking that down, RFC 3904 covers 4 cases described in the intro:
A) a gateway which does not provide IPv6 at all;
B) a dual-stack gateway connected to a dual-stack ISP;
C) a dual-stack gateway connected to an IPv4-only ISP; and
D) a gateway connected to an IPv6-only ISP.
In homenet, we are focused on case B and D. Specifically, that means we are not
going out of our way to ensure that the home network works well with Teredo,
6to4, and other transition mechanisms detailed in the RFC 3904. Also, within B,
we are charted to make IPv6 work without breaking IPv4 in the process, but I
think we can avoid defining ways of making sure IPv6 only hosts can talk with
IPv4 only hosts (that's behave's job anyway).
While in some cases the document goes in places we are not, in other areas it
doesn't go far enough. For example, Section 4 talks about meeting Case B (and
by reference later, D) requirements. One method described is extending a subnet
to span multiple links via ND Proxy. In contrast to this, one of the primary
goals of the homenet charter is to provide an agreed upon and simple to use
method of routing and prefix configuration among multiple subnets. If we are
successful (and ISPs avoid delivering service with a single /64 as well!) then
we may not need the ND proxy type of function at all.
- Mark
>
> I will be also trying to come up with more specific comments soon.
>
>
> Daniel
> --
> Soohong Daniel Park
> Samsung Electronics, DMC R&D
> http://www.soohongp.com, twitter:@natpt
>
>
> On Wed, Sep 21, 2011 at 9:40 PM, Tim Chown <[email protected]> wrote:
>
> On 19 Sep 2011, at 22:01, Mark Townsley wrote:
> >
> > Procedurally, the WG can do what it wants to here, there is no official
> > method for declaring a document a WG document. Different WGs operate in
> > very different manners in this regard.
> >
> > What I am suggesting here is that a draft should be published any moment
> > now (Tim?), and that I want people to read it and let us know if it is an
> > OK start for a work in progress for the WG to actively try to finish. If
> > there are no other competing architecture documents to evaluate, I'd like
> > to see it move towards being a WG document rather quickly (we're supposed
> > to be finished by December). This by no means says the document is finished
> > or even has consensus of the WG in its current state, just that it is the
> > document the WG will try and turn into a final product. All this means is
> > that it becomes the "working group's document" rather than "Tim, Jari,
> > Jason and Ole's document".
>
>
> Sounds fine.
>
> Now that the update is out, I would just say that so long as the WG knows
> this is "the homenet architecture draft", it's status is not that important,
> except that the charter target was WG adoption by the end of this month, and
> we shouldn't treat those targets too lightly.
>
> From the perspective of the interim meeting, I believe the agenda of that
> meeting includes slots for each of the five requirements (routing, naming,
> security, etc) so we should be agreeing as much as we can at the very least
> on the general principles we will follow for those slots, if not the
> architecture in full.
>
> Tim
> _______________________________________________
> homenet mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/homenet
>
>
>
> _______________________________________________
> homenet mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/homenet
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet
