RE: [Ecrit] Forest guides

["Brian Rosen" <br at brianrosen.net>]

> > Since we can't get a partial mesh, we may need some referral
> > operation where
> > an entity in the mesh undertakes to return a tree he is not
> > authoritative
> > for, but merely acts as an agent.
> >
> 
> That's already possible with the existing spec. Any LoST server can
> always return a referral (another LoST URL) in a mapping response or
> a special redirection error response. Hopefully, the details will be
> clearer in LoST -03.
That's pretty much what I was expecting, although it's hard to see how that
works for trees in the documentation as it exists today.  I'll wait for
Lost-03 and if I still get confused, maybe I'll suggest specifics.

> 
> 
> > I do think we want to have some more specifics.  I'm not sure if the
> > specifics can be an IETF document, but that would be my
> > preference.  What I
> > hope for is that there is a normative document that describes
> > exactly how
> > meshes and partial meshes work (in a protocol sense).  Then I think
> > we can
> > draft a BCP that would guide creation of the mesh by countries.
> > The latter
> > document would be something we would solicit input from NENA, the
> > EU Expert
> > Group on Emergency Access, the appropriate Asian agencies, etc.  If
> > we got
> > positive responses from them, we may have a solution.  It's a BCP,
> > not a
> > treaty.
> 
> The lost-sync document describes this, although it can probably use
> more detail in places. It allows any set of LoST servers to
> synchronize their data, be that FGs or authoritative servers. This
> can be either a full mesh or some partial peering mesh that floods
> mapping information by more than one step. Thus, I'd appreciate
> comments on the lost-sync document to make sure it has enough details
> to be usable.
I had not realized that this was the intended mechanism.  Again probably not
thinking of this as applying to the forest.  I'll review the documents
again.
> 
> I'm trying to minimize the number of mechanisms across the architecture.
Yeah, you've said that, and I understand it, but maybe we haven't gotten the
documents there yet.

Now that would be the protocol part.  A BCP on how you use it may still be
usefull.


_______________________________________________
Ecrit mailing list
Ecrit at ietf.org
https://www1.ietf.org/mailman/listinfo/ecrit