On Mon, Mar 3, 2014 at 1:07 PM, Joe Abley <[email protected]> wrote:
>
> On 3 Mar 2014, at 9:51, joel jaeggli <[email protected]> wrote:
>
>> On 3/3/14, 9:25 AM, Norbert Bollow wrote:
>>> Warren makes a strong argument in favor of .alt I think.
>>
>> yeah... anything that has the potential to result in additional leakage
>> seems like a recipe for additional pain.
>
> Well, except that the current proposal is to reserve (not delegate) ALT.
>
Weeeeeelll....
Actually it says (Section 3):
1. Stub resolvers MAY elect not to send queries to any upstream
resolver for names in the ALT TLD.
2. Iterative resolvers SHOULD follow the advice in [RFC6303],
Section 3.
3. The root zone nameservers should either return NXDOMAIN
responses, or the ALT TLD should be delegated to "new style"
AS112 nameservers. (TODO(WK): WK, JA, BD to revive AS112 /
AS112-bis).
Item 3 is specifically about this question -- it can either be that
the root continues to not know about the ALT "TLD"[0] or it could be
delegated to a new style AS112, which will, in theory, happily sink
$whatever.
That's an open question, but (IMO) a detail.
W
[0]: Much of this draft and discussion is made complicated by
terminology problems. If someone uses www.foo.tld in their own
protocol, it the rightmost label a TLD? Probably not... But, if the
name (which is *not* a DNS name), but is "DNS like" leaks into the
DNS, then it is...
> If we assume that leaks will happen, then they will hit the root servers and
> there's no opportunity to sink the queries anywhere else.
>
> If we delegate ALT, then we have to decide where to. I can see this being
> contentious.
>
>
> Joe
_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop
