Re: [Add] The ADD WG has placed draft-schwartz-svcb-dns in state "Call For Adoption By WG Issued"

[Ben Schwartz <[e-mail]>]

On Mon, Sep 27, 2021 at 6:59 PM Brian Dickson <[e-mail]>
wrote:

> I just read the companion drafts in ADD.
>
> I think my concerns would be mitigated with the following changes:
>
>    - Explicitly limit the use case to recursive server(s)
>    - Change the binding to a different mnemonic/label that does not squat
>    on the "dns" label which may be more applicable to an authoritative server.
>    I'd suggest "resolver", as it is consistent with usage from the DNS
>    terminology RFC (but I haven't looked at that lately, so I may be
>    misremembering the terminology).
>
> As long as this does not eventually conflict with an SVCB binding for a
> DNS authoritative server (something I'm working on in DNSOP), I think it
> would be okay to adopt in ADD.
>

Hi Brian,

The binding here is deliberately independent of recursive vs. authoritative
roles.  It is simply a binding for the DNS protocol.  Recursive vs.
authoritative is not strictly a property of a DNS server (or client);
rather, it is a property of an individual message (RD=0/1, AA=0/1).  In
some cases, a single server may offer both functions, or they may be mixed
in some fashion.  For example, there is a draft currently under
consideration in ADD [1] that uses a hybrid approach to identify resolvers
that are also authoritative for certain domains.

Also, many codebases combine stub, recursive, and authoritative
functionality.  Providing a unified mapping, independent of role,
simplifies implementation by improving separation between transport setup
and message contents.

I'm not aware of any conflict between this draft and any of the active
proposals in DNSOP or DPRIVE.  However, if we discover a need for further
SvcParams in some DNS use case, we can certainly define them, either in
this document or a subsequent one.

--Ben Schwartz

[1]
https://datatracker.ietf.org/doc/html/draft-reddy-add-enterprise-split-dns-06

>