Forgot to mention: this thread is based on an open Github issue - 
https://github.com/anima-wg/constrained-voucher/issues/281 
This checks whether cBRSKI can be extended to new cert formats if those become 
popular. That seems ok, using either option 1 or 2.

> So either way we need to do something now in RFC8366-bis to avoid rolling
> that module again later on.

With option 2, we can adapt the RFC8366-bis YANG descriptions now, to prepare 
for new/other cert and key types in the future.
With option 1, we can just do nothing now but it would require an update of 
RFC8366-bis again for each new cert/key type that gets added.

So option 2 is some work (now). Certain YANG field descriptions need to be 
generalized to allow the (future) new key and cert types.

> Are you interested in it?
> Are you writing code?

For now and the coming year, I don't have any plans to use C509 certs or write 
code for it! If no-one has such plans, we can also decide to go for option 1.

> I was hoping (my head in the sand) you wouldn't bring this up :-)

I was digging in that sand too deep and found something :)  If we treat support 
for a new cert format (in the DTLS handshake, and in the Voucher) as simply a 
"new Voucher format" , we don't have to do anything special for this case.
Nothing on top of the things already defined now in 
draft-eckert-brski-discovery. The "new Voucher format" could even have its own 
media type to distinguish its use of different cert/key encoding.

Esko

-----Original Message-----
From: Michael Richardson <[email protected]> 
Sent: Thursday, December 14, 2023 14:38
To: Esko Dijk <[email protected]>; [email protected]
Subject: Re: [Anima] Voucher RFC8366-bis: support for other types/encodings of 
certificates?


Esko Dijk <[email protected]> wrote:
    > How would the voucher / voucher-request best support such new
    > certificate formats/types? Two main approaches are possible:

Yeah.
So either way we need to do something now in RFC8366-bis to avoid rolling
that module again later on.

    > So with option (2) the existing fields just hold binary data of some
    > format, by default it's X509/ASN.1 DER. If the "cert-type" field is
    > present it identifies the format of binary data e.g. X509, C509, etc.

    > If we do nothing now, we will default to option 1. TBH I'm not even
    > sure if option 2 is feasible with current YANG definitions and all the
    > legacy around it. But I still wanted to raise this question, just in
    > case anyone's interested :-)

Are you interested in it?
Are you writing code?

    > PS Also for discovery operations, discovering a Registrar that supports
    > a particular (deviating) certificate type X may then be needed. This
    > could be viewed as just a different type of Voucher that needs to be
    > supported.

I was hoping (my head in the sand) you wouldn't bring this up :-)


--
Michael Richardson <[email protected]>   . o O ( IPv6 IøT consulting )
           Sandelman Software Works Inc, Ottawa and Worldwide




_______________________________________________
Anima mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/anima

Reply via email to