DNSOP minutes from IETF 56

[Rob Austein <sra+dnsop@hactrn.net>]

It looks like the DNSOP minutes from IETF 56 were sent to the list but
never arrived (sigh), so here they are (again).

========

Domain Name System Operations (DNSOP)

Chairs:		Dave Meyer and Rob Austein
Minutes:	Sean McCreary (with minor edits by by chairs)

Dave introduced himself and Rob Austein, the new DNSOP chairs.  Dave
presented the agenda.  Agenda very tight due to short meeting slot.
Sean McCreary kindly volunteered to take notes.

========

        Alain Durand presented draft-ietf-dnsop-ipv6-dns-issues-02.txt
<http://www.ietf.org/internet-drafts/draft-ietf-dnsop-ipv6-dns-issues-02.txt>
        "IPv6 DNS transition issues"

Rob Austein:    Draft discusses two issues
        Dual stacking (we had consensus on coexistence plan in previous mtgs);
        Stuff about the reverse tree.
                Would it be useful to split these two into separate drafts?

        Rob asked for a hum.
                Clear consensus was for a split.

Alain asked for ISPs to send comments to the mailing list.

========

Randy Bush presented DNS issues for 6to4

        draft-moore-6to4-dns-03.txt
        6to4 delegations should come from same place as IPv4 reverse
                delegations

        What if upstream refuses to deal w/v6?  Then go up the tree
        towards the root until you reach someone who cares.
        This will force the registries to do direct delegations to end sites.
        draft-ymbk-6to-4-arpa-delegation-00.txt

Alain Durand:   What about delegations for dynamically allocated addresses?

Randy "Let's not create problems for a simple hack that works."

Dave called for a hum.
	No consensus in room for adopting these drafts as working
        group documents.  Discussion to continue on mailing list.

================

Matt Larson asked if draft-ietf-dnsop-bad-dns-res-00.txt should be
resubmitted, as it has expired.

Rob: It should be if people are willing to work on it, and asked for a
hum.  Appeared to be consensus to continue work on this document.

Matt will resubmit it.

==============

Akira Kato presented draft-kato-dnsop-local-zones-00.txt, "Operational
Guidelines For `Local' DNS Zones".  The pseudo-TLDs ".local" and
".localhost" are generating significant load on the root servers.

Rob Austein asked if Apple's "Rendezvous" uses port 53.

Stuart Cheshire said rendezvous uses port 5353 for queries.

Geoff Huston asked whether the proposal to "delegate" .local to AS112
servers means creating a new TLD.

Paul Vixie: We propose to add a top level domain, but a useless one
that always returns NXDOMAIN.  The only purpose is to reduce the load
on the root servers.

Stuart: This can be solved in the resolver by filtering queries to the
.local pseudo-TLD, or by changing the root infrastructure.

Topic too new and complex for immediate hum on whether WG should take
this on.  Discussion to continue on mailing list.

==============

Paul Vixie spoke on a proto-draft he had posted to the mailing list on
size issues in DNS response messages.

Paul's analysis suggests that the naming scheme used by the name
servers for the root zone (x.root-servers.net) saves a lot of space in
response packets, which increases the number of name name servers that
can be listed for a zone.  Paul suggests that we should encourage
others to adopt this scheme.

If the IDN effort increases the size of the A record, reducing the
number that will fit in a response packet.  How many A records are too
few?  Is only one too few?  If one is an acceptable number, what kind
of engineering changes can be made so that the A record returned is
useful to the client.

If this is relevant to the working group, we should adopt it and
advance it as soon as possible.

Discussion taken to mailing list.

========================

Johan Ihren presented draft-ihren-dnsext-threshold-validation-00.txt
<http://www.autonomica.se/~johani/talks/ietf56-threshold-validation.pdf>
        A discussion of how to use DNSSEC key signing keys (KSKs) to
        improve reliability in case some private keys are compromised,
        and during key rollovers.

Johan then presented draft-ietf-dnsop-interim-signed-root-01.txt
This is a clarification of the earlier draft.

Johan asked if threshold validation should be taken as a work item for the
        working group (as a general topic, not just Johan's draft).
        Consensus was yes.

=======================

Rob asked if the working group should produce a requirements draft for
        Autoconf/bootstrap/dynamic update/dns discovery with IPv6 clients.
        Rob believes the work group should work on this.

Alain Durand:  Does this make more sense to work on in a working group
        dedicated to finding a general solution to this problem?

Rob:    There are DNS-specific issues that need to be identified.

Stuart Cheshire: Determining DNS parameters is a major hole in current
        IPv6 autoconfig protocols.

Bob Hinden agreed with Stuart, and added that once a scheme is defined and
        deployed it will be possible to to a lot more with autoconfiguration
        for v6 clients .

Erik Nordmark:
	I've tried to raise interest in solving the generic problem and not
        succeeded.  There is interest in solving the specific part of the
        problem for DNS.

Mike Patton:  DNS itself is a leveraging thing that is really needed, other
        complications will produce a heavyweight solution.
        People will want a lightweight solution anyway.

Rob called for a hum.
        Consensus was that the working group should work on this.

===========================

Alain Durand presented draft-durand-dnsop-dynreverse-00.txt

He asked that this be adopted as a working group document.
        No consensus for or against.
Rob deferred a decision until there is more discussion on the mailing list.

=============================

Jae-Hoon Jeong presented draft-jeong-hmipv6-dns-optimization-00.txt
        A scheme for supplying DNS services in a mobile IP environment.

Rob deferred adoption of the document to a mailing list discussion.

=============================

draft-warnicke-network-dns-resolution-00.txt
        An update for RFC1101, a scheme for finding the default gateway
        for a client's local LAN.

Rob:  This is a request for review of the propsal, not a request to for
	the working group to adopt the document as a work item.

============================

As the previous speakers had run over, Rob and Dave deferred the
charter discussion that had been on the agenda.  Charter discussion
will take place on the mailing list.
#----------------------------------------------------------------------
# To unsubscribe, send a message to <dnsop-request@cafax.se>.