On Thu, Jun 24, 2021 at 2:38 PM John Levine <[email protected]> wrote:
>
> It appears that Ben Schwartz  <[email protected]> said:
> >I think the "Privacy Considerations" section should probably mention QNAME
> >minimization, which ought to help a little.
>
> I'd also like it to say more clearly up front that .ALT is for names that are
> totally outside the DNS protocols, not for names handled locally using DNS 
> protocols.
> It's for things like .onion, not like .local.


Yup, what John said.

The important part is: "This document reserves a string (ALT) to be
used as a TLD label **in non-DNS contexts.**" (emphasis added).

There are a bunch of different name resolution mechanisms - DNS is the
most common, but there are also an increasing number of things like
.onion, various blockchain-based systems (some of which have some cool
privacy features), etc.
For obvious reasons "alternate"[0] resolution systems use the same
format/style of names as the DNS (LDH, case insensitivity, etc) -
people want to be able to enter the identifiers into "normal"
applications and have them just work (e.g things like "foo:bar:baz"
don't work - ssh[1] for example won't pass this to a resolution call).
There are similar issues with defining new schemas, etc; there is a
reason that it's e.g http://www.facebookcorewwwi.onion/ and not
http://www.facebookcorewwwi.¯\_(ツ)_/¯/ or
hQQp://www.facebookcorewwwi.onion/.

Unfortunately, having multiple resolution systems using the same
namespace and syntax does not provide a signal to denote which
resolution mechanism should be used - clearly .com is "in the DNS" and
.onion isn't -- but this doesn't scale, and simply saying "the DNS is
the only resolution system" doesn't either....

I'm guessing that everyone knows all this, but we had lots of *long*
discussions around the various options in ~2014 - 2017, and I figured
that much of the state might have been paged out...
Some other background:
https://www.iab.org/activities/workshops/explicit-internet-naming-systems-ename/
https://www.ietf.org/blog/ename-workshop/


>
>
> >I would also be interested in seeing some guidance about interaction
> >between the relative form (.alt) and good old-fashioned search domains.  It
> >seems to me that the interaction there is poor... perhaps bad enough to
> >recommend using the absolute form only.
>
> I thought we all agreed that search lists are bad when .CS was added to the 
> root.


Yup. There should indeed be some more text around search-lists and
search-list processing; much of the purpose of this is to be able to
differentiate the namespaces, and search-lists are a place where they
may bump into each other again.
John's right again -- we *did* all agree that search lists are bad
when .CS was added... and also in RFC 1123 (specifically, section
6.1.4.3 (2)), RFC 1535, and RFC 1536.
Andrew McConachie and I even started writing "DNS Search Lists
Considered Dangerous", before deciding that this was probably tilting
at windmills...


W
[0]: This isn't intended to be pejorative, I'm just trying to separate
these from what we normally assume in this community when we talk
about a name resolution system
[1]: OpenSSH_8.6p1, OpenSSL 1.1.1k  25 Mar 2021


>
>
> R's,
> John
>
> _______________________________________________
> DNSOP mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/dnsop



-- 
The computing scientist’s main challenge is not to get confused by the
complexities of his own making.
  -- E. W. Dijkstra

_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to