[I know there is a post-Yokohama version in construction, addressing,
among other things, the authentication issues which will move to a
separate I-D, and the questionable fragmentation which will
disappear. Still, these points may be useful:]

On Mon, Oct 19, 2015 at 03:10:50AM +0000,
 Tirumaleswar Reddy (tireddy) <[email protected]> wrote 
 a message of 70 lines which said:

> This revision addresses comments from the WG. Main changes are:

> DNS privacy problem is further discussed in
> [I-D.bortzmeyer-dnsop-dns-privacy].

Now RFC 7626

> The existing Query ID allows multiple requests and responses to be
> interleaved in whatever order they can be fulfilled by the DNS
> server. [...] When sending multiple queries over a single DTLS
> session, clients MUST take care to avoid Message ID collisions.  In
> other words, they MUST not re-use the DNS Message ID of an in-flight
> query.

Since DNS-over-DTLS is, in that respect, closer from DNS-over-TCP than
from DNS-over-UDP, would it make sense to simply refer to 5966bis,
sections 6.2.1 and 7?

> It is highly advantageous to avoid server-side DTLS state and reduce
> the number of new DTLS sessions on the server which can be done with
> [RFC5077].

Since TLS session resumption is not ideal for privacy, may be
referring to the analysis of RFC 6973, section 5.2.1 would help?

> Implementing DNSoD on root servers is outside the scope of this
> document.

s/root/authoritative/ ?

> 14.1.  Normative References

The list of normative references is awfully long and not all RFCs in
it seem worth it (for instance RFC 7469 or 7435). I suggest to move
most of them to informative.

_______________________________________________
dns-privacy mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dns-privacy

Reply via email to