CURRENT_MEETING_REPORT_

Reported by Randy Bush/RGnet

Minutes of the DNS IXFR, Notification, and Dynamic Update
Working Group (DNSIND)

Thanks to Kent Malave for acting as scribe for the meeting.


Administrivia

This is hoped to be the last meeting of the DNSIND Working Group.  The
group plans to close after the drafts move to Proposed Standards.


Dynamic Update Draft (draft-ietf-dnsind-dynDNS-02.txt)

Changes in the latest draft:

   o Different message format.

   o Why?  Need extra section for zone, prudent for extensibility, and
     reflects intended semantics in full generality.

   o Header:  first eight bytes of query header.

   o Body:  sequence of self-encoded sections.

   o Section:  code, length, value (in RR's).

   o Five Sections:  ZoneName, ADD, DELETE, ADDNAMENEW, and
     ADDNAMEEXITS.


The following comments were made in the meeting:

   o Serialization needs to be stated as optional in the draft.

   o Any new changes to the draft need to be shipped soon.

   o IBM announced that it will make available changes to the public
     implementation of BIND to implement dynamic update.

   o A potential bug in the glue area code was presented.  The
     resolution of this is that if the ZoneName section is used, there
     is no bug.

   o There were objections to overloading the TTL vs.  length for
     wildcards.

   o There were objections to overloading sig record.

   o Nameserver forwarding requests to the slave needs a time-out.

   o Will ixfr and dynup have the same packet format?  It is nice,
     interesting, but not necessary.



Incremental Transfer Draft (draft-ietf-dnsind-ixfr-02.txt)

Changes and additions to the draft include a suggestion has been added
to ignore responses without UDP checksum and examples have been fixed.

The following comments were made in the meeting:

   o There is potential for a problem if the zone is changed but the
     serial is not updated.  A potential resolution is to send back a
     hash along with the serial to track the changes.

   o In Section 5, it may be nice to state that information ``can'' be
     purged.


A two-week preliminary last call will be followed by the formal Last
Call.



Notify Draft (draft-ietf-dnsind-notify-01.txt)

The following comments were made in the meeting:

   o Code does not match the draft.  The code is being updated to match
     the draft.

   o Are version fields needed?  Zero fields should be ignored on
     current implementations.

   o Why just Internet class?  SOA spans classes so this will be looked
     at by the author.

   o Should MUST, SHOULD and MUST be capitalized?  The point was dropped
     so no changes are needed to the draft


A two-week preliminary last call will be followed by the formal Last
Call.


Classless in-addr.arpa Draft

The document was presented:  reasoning defined, examples given,
solutions proposed, and experiences given.  It was noted that simple
examples are needed for this draft.

The document will be moved to Informational RFC or BCP RFC.