Re: [Add] Resolver Information Discussion

[Vittorio Bertola <[e-mail]>]

> Il 05/05/2022 20:32 Deen, Glenn <glenn_deen=[e-mail]> ha scritto:
> 
> 
> Has the base design gone in a different direction than originally considered to the point where extended information about resolvers is no longer of interest to clients? 
> 
> Is Resolver Information an area that is still of interest to the group?
> 
The desire to address this topic was prompted by the fact that early implementations of encrypted DNS put policy requirements on a resolver's behaviour before accepting to use it, so it would be necessary for the local resolver to state its policy to the browser so that it could make a decision. Later implementations indeed went in a different direction, i.e. using DDR to upgrade the existing resolver no matter which policies were in place, and perhaps in the future using DNR to get the local encrypted resolver from the network.

However the original approach is still there in at least one browser (Firefox) and it is unclear what they plan for the future. It seems to me that they would be the primary consumers of this information, but I have no idea if they would be interested in evolving their TRR model to an open, discovery-based mechanism or not. Indeed, they'd have the problem of deciding how much one can trust a resolver's self declaration of policies.

--

Vittorio Bertola | Head of Policy & Innovation, Open-Xchange
[e-mail] mailto:[e-mail]
Office @ Via Treviso 12, 10144 Torino, Italy