Ok, I went back and read parts of RFC 8995 (Section 5.6, 11.4, other parts
of Section 5), and a tiny bit of RFC8366.  I also watched mcr's video
(which does explain it nicely, x1.5 is the bomb).  To clarify my concern
(to David), I was discussing exactly what is mentioned in 8995's section
11.4 (the name of that section doesn't help).  So, my concern is addressed.

I do think adding the voucher request/response to Figure 2 would help.  But
I fear it would make that figure much more complicated (although it just
adds a column?).

I also see now that if someone wants to use CMP (or SCEP) they will also
have to deal with EST/CMS for parts of the process.  I wonder just how
complex these devices will have to be.  But that isn't really a comment
against the draft.

Thanks for taking the time to explain.  I'll re-write my discuss as a 'no
objection' w/ a comment with some part of the above.

Deb



On Mon, Sep 2, 2024 at 5:01 PM David von Oheimb <
[email protected]> wrote:

> s/need/do not need/
> On 02.09.24 22:58, Michael Richardson wrote:
>
> Deb Cooley <[email protected]> <[email protected]> wrote:
>     > If there is always a TLS/DTLS session, then why is there a note in Sec
>     > Consid about CMP messages being in the clear?
>
> 1. brski-ae also can use brski-prm [they were one document before].
>    PRM must use CMP, as it uses HTTP between a new thing (the
>    "registrar-agent"), and the pledge.  With the initiation reversed.
>
> 2. the connection between RA and CA might not be encrypted.
>
> In general, CMP messages are designed so that they need transport security.
> (vs EST)
>
>
>
> --
> Michael Richardson <[email protected]> <[email protected]>   . o O ( 
> IPv6 IøT consulting )
>            Sandelman Software Works Inc, Ottawa and Worldwide
>
>
>
>
>
>
_______________________________________________
Anima mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to