Re: [dnsext] afasterinternet.com trial and draft-vandergaast-edns-client-subnet-00

[Ted Hardie <[e-mail]>]

On Mon, Sep 5, 2011 at 10:09 AM, Wilmer van der Gaast <[e-mail]> wrote:
> On 31 August 2011 17:03, Ted Hardie <[e-mail]> wrote:
>> * Rewrote origination sections to clarify that normally only recursive
>> resolvers generate edns-client-subnet options.
>> * Discussion on whitelisting or automatically detecting if an
>> authority supports edns-client-subnet.
>> * More precisely specifying (optional) transitive behaviour.
>> * Minor revisions to improve clarity and correctness w.r.t. other RFCs.
>>
>> None of these changes seem to address the balance of discussion of the
>> previous two versions.
>>
> I assume you are referring to the discussion about edns-client-subnet
> or DNS-based CDNs being a reasonable solution at all?

Not exactly, no.  Some of those discussions suggested that this be opt-in
rather than opt-out (so that the EDNS option is set by the client when
desired, rather than it setting 0.0.0.0. when it does not desire it).

The basic issue here is that this effectively changes the q-tuple to include
a subnet as part of the question being asked.  A discussion of why
an intermediate should be the one to change the question seems warranted,
if only to highlight the parameters  of the experiment.

regards,

Ted Hardie