How would we use DNR?

[cabo]

DHCP or Router Advertisement options [I-D.ietf-add-dnr]

I'm not sure this is defined yet -- is it intended to be?

[miri64]

The draft for that is currently discussed (last call) in the add working group: https://datatracker.ietf.org/doc/draft-ietf-add-dnr/

[cabo]

Yes, you reference that.
Does that draft explain how to refer to DoC?

[miri64]

Does that draft explain how to refer to DoC?

Indeed, there seems to be some extra work required to make this work with DoC. From a quick read-through (specifically Section 3.1.5 of -13 of that draft) I'd say at least an ALPN needs to be defined (though how to do that with OSCORE or if it is even required in that case, I am not sure). The port should be a given.

Maybe this option is only suitable if one chooses to use CoAPS, as it seems very tailored to TLS and for OSCORE (or unencrypted CoAP use) we need another way (e.g. as mentioned a resource directory).

[miri64]

After we resolved the ALPN issue in DNRs https://mailarchive.ietf.org/arch/msg/core/yFHfHUHDDwh0HANDJEqj_TmkiMc/, there is still the question on how to represent DoC (or CoAP services in general) in a SVCB record (the format of which would provide the DNR Service Parameters as mentioned in https://datatracker.ietf.org/doc/draft-ietf-add-dnr#section-3.1.5).

To summarize the findings: While a coap ALPN exists already [RFC 8323],

• it is only ever mandated for TLS, not DTLS https://mailarchive.ietf.org/arch/msg/core/I_vkEz046qEWQKGh6dOE1WobIM4/,
• just from ALPN and port alone we can not distinguish between coaps, coaps+tcp or even coaps+ws (not sure this one is needed for SVCB though) or any other transport for that matter,
• there is no SVCB definition for OSCORE or EDHOC (https://mailarchive.ietf.org/arch/msg/core/Bc36wzeg7bzNxNgj4W1k7GEErKw/)

Given all that, and the fact that this also may effect CoAP services as a whole, not just DoC, I see the potential for further draft(s) that would define the Service Parameters for the existing CoAP transports, that we than need to reference in what is currently Section 3 of the DoC draft.

Do you agree?

[EskoDijk]

We have now the DDR / DNR RFCs:

• https://datatracker.ietf.org/doc/html/rfc9462
• https://datatracker.ietf.org/doc/html/rfc9463

I was also wondering if these should be referenced and how discovery of DoC service might be done using such mechanisms.

[miri64]

#27

[miri64]

There is now https://datatracker.ietf.org/doc/draft-lenders-core-dnr/, which is referenced as of -06 of the DoC draft.