Re: [dns-privacy] ADoX experiments (was: Re: Intermediate proposal (what I was saying at the mic))

[Stephen Farrell <[e-mail]>]

Hiya,

On 17/08/2021 13:16, Brian Haberman wrote:
> All,
>       I want to start working through the details of what Stephen is
> proposing below; beginning to do operational experiments to determine
> which approach, or approaches, may be viable in the long term. To carry
> out such experiments, I believe we need the following:
> 
> 1. A stable I-D for an approach to providing privacy between recursive
> resolvers and authoritative servers,

More than one I-D seems likely and is fine I reckon.

> 
> 2. An implementation of the stable I-D in a recursive resolver and in an
> authoritative server implementation,

IIUC, that exists?

> 
> 3. At least one authoritative server operator willing to deploy the
> experimental implementation,
> 
> 4. At least one recursive resolver operator willing to deploy the
> experimental implementation,
> 
> 5. An agreed upon set of metrics to assess the operational behavior of
> the approach,

Not sure a final/agreed version of that's needed before
someone starts to run an experiment. Nice to have it of
course but I suspect initial experiments will throw up
bits of data we'd not considered (e.g. maybe relating to
the diversity of partnerships that authoritative servers
have and how those affect partial rollout).

> 
> Is there a major item missing from the list above? Other aspects of
> carrying out such an experiment?
> 
> Are there any volunteers to start working on details of such an experiment?

I'm game. (But not this week, taking a few days away:-)

Cheers,
S.

> 
> Regards,
> Brian
> 
> On 8/2/21 9:22 AM, Stephen Farrell wrote:
>>
>> Hiya,
>>
>> On 02/08/2021 05:21, Martin Thomson wrote:
>>> If we decided on a single answer for the first and in the negative
>>> for the second, would that make authentication viable?
>>
>> IMO we ought not just "decide" on most of the tricky ADoX
>> issues but we should rather document the options sufficient
>> to allow people to do experiments and then wait and see how
>> those experiments go. I'd say a stable I-D is probably
>> enough documentation to allow for experiments and I'd hope
>> such experiments could be done in 6-12 months. I'd expect
>> we might still be left with a few tricky issues, but that
>> a number of those (where we might make wrong choices now)
>> would be resolved once people try 'em out.
>>
>> So my suggestion is to review the I-Ds we have with a view
>> to figuring out what's missing that's needed to allow such
>> experiments, fix that and then "park" those I-Ds 'till we
>> get results. That could be similar to how drafts are declared
>> to be "interop drafts" in other WGs or could be a WGLC-like
>> process.
>>
>> If we decided to try go that way, I'd be happy to try help
>> get some such experiment going.
>>
>> Cheers,
>> S.
>>
>> _______________________________________________
>> dns-privacy mailing list
>> [e-mail]
>> https://www.ietf.org/mailman/listinfo/dns-privacy
>>
> 
> 
> _______________________________________________
> dns-privacy mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/dns-privacy
>