> On Mar 18, 2021, at 9:40 AM, Eric Orth <[email protected]> 
> wrote:
> 
> 
> 
> On Thu, Mar 18, 2021 at 12:33 PM Jim Reid <[email protected] 
> <mailto:[email protected]>> wrote:
> 
> 
> > On 18 Mar 2021, at 16:21, Eric Orth <[email protected] 
> > <mailto:[email protected]>> wrote:
> > 
> > I disagree with your assumption that clients/users are only concerned about 
> > particular resolvers. 
> 
> Eric, I didn’t make any assumptions about that at all. It was Tommy who said 
> ODNS would benefit those who were concerned about leakage to very large 
> public resolvers. All I did was suggest a simpler alternative that wouldn’t 
> need an RFC or the introduction of more complexity and lots more moving parts.
> Ah, thanks for the explanation.  Maybe Tommy can further explain his point?
>  
> 
> If the aim of ODNS is to prevent leakage to resolvers *in general*, then 
> that’s a different story. But as I said the use case and problem statement 
> isn’t (yet) compelling enough. The cost/benefit analysis is unclear too. YMMV.
> Preventing leakage to resolvers *in general* would be exactly my interest in 
> ODoH.  This goes along with my previous statement that I think clients could 
> find good use in automatically using ODoH (for either particularly-sensitive 
> or all queries) as an autoupgrade to known-supporting providers rather than 
> just being something users explicitly select when they have particular 
> concerns.

To clarify my point earlier, I was mainly replying to the concern that 
supporting ODoH would be a burden for resolvers to implement. Although I don’t 
think it’s actually that much complexity to add, given our experience, I don’t 
see a deployment scenario where all recursive resolvers would offer oblivious 
resolution.

Mainly, this is relevant for resolvers that directly receive client-generated 
queries, are not directly on the local network (and thus already know a lot 
about the user), and are in a position where a client can use a proxy to access 
them. Large public resolvers do fall into this category, but it isn’t limited. 
It could also be the resolvers used for ISPs, carriers, etc.

Thanks,
Tommy

>  
> 
> 
> _______________________________________________
> dns-privacy mailing list
> [email protected] <mailto:[email protected]>
> https://www.ietf.org/mailman/listinfo/dns-privacy 
> <https://www.ietf.org/mailman/listinfo/dns-privacy>
> _______________________________________________
> dns-privacy mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/dns-privacy

_______________________________________________
dns-privacy mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dns-privacy

Reply via email to