Re: [DNSOP] comments on dnsop-qname-minimisation-02

[Shumon Huque <[e-mail]>]

On Wed, Mar 11, 2015 at 12:50 PM, Paul Hoffman <[e-mail]>
wrote:

> On Mar 11, 2015, at 9:02 AM, Stephane Bortzmeyer <[e-mail]>
> wrote:
> >
> > On Wed, Mar 11, 2015 at 12:35:29AM -0400,
> > Shumon Huque <[e-mail]> wrote
> > a message of 400 lines which said:
> >
> >> Are we standardizing on the british spelling of "minimisation" in
> >> preference to the americanized "minimization"?
> >
> > Bikeshedding is postponed until Working Group Last Call :-)
>
> Or beyond. The RFC Editor allows both types of spelling, and they will
> make it consistent.
>

Okay, that's fair!


> >
> >> One thing this document doesn't make clear is that the algorithm
> >> being presented not only minimizes the query name, but also hides
> >> the query type until it reaches the target zone (by using the NS
> >> query type rather than the actual type).
> >
> > Do note the use of NS is not mandatory. See section 3, the paragraph
> > starting with "Another way to deal with such broken name servers"
> > (which you mention later) and also section 3, 1st paragraph about the
> > statistics of qtypes.
>
> My strong preference is that this document only deal with qname
> minimization. If someone wants to extend that to qtype minimization, which
> covers a different threat model, that should be done in a different
> document.
>

I'm not yet fully decided on this, but I think I'm leaning in Paul's
direction. By deferring
the qtype hiding feature, we have a simpler resolution algorithm & smaller
changes to
existing code. And perhaps more importantly, we have easier deployability -
we don't
have to implement a specific set of workarounds for known problems such as
authority
servers not responding to NS queries, or middleboxes that don't allow them.
And we save
the wasted time & cycles involved in executing those workarounds.

Shumon Huque