Re: [Add] Designated Resolver Terminology

[Tommy Jensen <[e-mail]>]

Hey Jim,

If I am following your proposal correctly, you are saying change “designated resolver” within DDR to be “designated caching resolver” so that if/when Adaptive or any similar draft defines the concept of a resolver designated for a set of names, it can be called “designated authoritative resolver” without needing at that time to refine names with DDR.

Is that correct? Or do you see the different cases within DDR as needing two distinct names?

Thanks,
Tommy

From: Add <[e-mail]> On Behalf Of Jim Mozley
Sent: Thursday, March 25, 2021 4:20 AM
To: [e-mail]; Tommy Pauly <[e-mail]>
Subject: [EXTERNAL] Re: [Add] Designated Resolver Terminology

Hi ADD,

Following up on my own post below. Given that changes are being made to draft-ietf-add-ddr would it be appropriate to consider the following terminology, or something similar, to disambiguate designated authoritative from caching resolvers:

“Designated Authoritative Resolver” – for designated authoritative resolvers
“Designated Caching Resolver” – for caching resolvers

There had also previously been a definition of “companion resolver” as well. I do realise a system could be both authoritative and caching (even if this is inadvisable), but I think two distinct names could help here in disambiguating the role being discussed in this case.

Over the history of DNS there has sometimes been terminology introduced that can be confusing, so it would be good to avoid confusion when introducing designated resolvers. What’s in a name? 😊

Jim

On 07/03/2021, 10:54, "Jim Mozley" <[e-mail]<mailto:[e-mail]>> wrote:

On 05/03/2021, 17:54, "Tommy Pauly" <[e-mail]<mailto:[e-mail]>> wrote:

Hi Jim,

Yes, there is a different use of the term between the documents. The adopted draft is trying to have the right terminology to describe the intent of the mechanism for discovering capabilities via SVCB records as described.

Fundamentally, the mechanism in draft-pauly-add-resolver-discovery is also discovering resolvers based off of SVCB records. The difference lies in which party is designating: a resolver designating another resolver, and a domain owner designating a resolver. I think we could start calling the latter case (in the document not adopted) "Domain-Designated Resolver” or something similar. We can update that some time.

Hi Tommy and ADD, I like “Domain-Designated Resolver” or similar for the authoritative server use case. I think it could be less ambiguous to establish a name for the resolver designating another resolver. I realise “Resolver-Designated Resolver” could be a mouthful. For consistency how about “Designated Domain Resolver” and “Designated Caching Resolver” as being consistent naming and dis-ambiguating?

Cheers, Jim