Re: [lamps] [Anima] /.well-known/brski reference to brski-registry

["Brockhaus, Hendrik" <[e-mail]>]

> Von: Benjamin Kaduk <[e-mail]>
> Gesendet: Dienstag, 5. April 2022 19:32
> 
> On Sun, Apr 03, 2022 at 10:36:01PM -0400, Sean Turner wrote:
> >
> >
> > > On Apr 1, 2022, at 02:25, Brockhaus, Hendrik
> <[e-mail]> wrote:
> > >
> > >
> > >> Von: Russ Housley <[e-mail]>
> > >> Gesendet: Donnerstag, 31. März 2022 19:53
> > >>
> > >>> On Mar 31, 2022, at 12:20 PM, Brockhaus, Hendrik
> > >> <[e-mail]> wrote:
> > >>>
> > >>> Thank you Michael for rising the questions.
> > >>>
> > >>>> Von: Anima <[e-mail]> Im Auftrag von Michael
> > >>>> Richardson
> > >>>> Gesendet: Donnerstag, 31. März 2022 17:48
> > >>>>
> > >>>>
> > >>>> We were discussing the /.well-known/cmp that is in being proposed
> > >>>> in
> > >>>> draft-ietf- lamps-cmp-updates, We were comparing it to
> > >>>> /.well-known/brski and /.well- known/est.
> > >>>>
> > >>>> Question 2)
> > >>>>  Should the CMP document be establishing a registry or not?
> > >>>>
> > >>> As discussed during IETF 113 I plan to do these things in CMP
> > >>> Updates
> > >>> - register 'cmp' in the "Well-Known URIs" registry
> > >>> - define a protocol registry group "Certificate Management Protocol
> (CMP)"
> > >>> - define a registry for "CMP Well-Known Arbitrary Label URI Segments"
> > >> defining 'p' to be followed by a <profileLabel>.
> > >>> In addition I would define a registry for "CMP Well-Known
> > >>> Operation Label URI
> > >> Segments" in Lightweight CMP Profile containing the path segments
> > >> defined three for http and coap use.
> > >>>
> > >>> Does this makes sense?
> > >>
> > >> Hendrik:
> > >>
> > >> That is consistent with the discussion lat week.
> > >>
> > >> Russ
> > >
> > > Would it also be sufficient to have only one additional registry "CMP Well-
> Known URI Path Segments" containing the arbitrary label 'p' and the operation
> labels?
> > >
> > > Hendrik
> >
> > When the /.well-known/est/ was registered we only did the top level, i.e.,
> /est/. There are no registries for the /.well-known/est/*this part*.  It’s not clear
> to me that you need to do anything more than get /.well-known/cmp.
> >
> > What will be the registration policy [0] for the ‘p’ values? I assume FCFS (first
> come first served)?
> 
> I had assumed that we were just registering the value 'p' in a single combined
> registry of CMP operations and path labels, but that the stuff after 'p' was site-
> local and did not need to be registered.  (Though a FCFS registry for them is not
> wrong.)
> 
> -Ben

Ben, thank you for clarifying this. This is also my understanding and I am about to update the drafts accordingly.

Hendrik