Minutes from RIPE 37

RIPE 37 DNS Working Group
14th September 2000, 9am

Chair: Randy Bush, Verio 
(standing in for Ruediger Volk)

Scribe: Lee Wilmot, RIPE NCC

Agenda: not distributed prior to meeting

Resulting action points: none

-------------------------------

1) IETF: DNSEXT WG

2) IETF: DNSOP WG
Randy Bush
    - root NS requirements draft through
    - thinking about DNSSEC deployment

3) IETF: Domain name internationalisation
Harald Alvestrand

  - handling non-ASCII domain names e.g. from non-English alphabets.
  - least pain decided from translations of foreign names into ASCII
    gibberish and back
  - some people starting to use non-standard domain
    names already, therefore some pressure to assert standard
    Many companies not bringing their proposal to the IETF
  - Harald hopes within 6 months a solution thats not too painful
    to users
    - main problem: how do apps handle these names bubbling up.

4) DNSSEC Workshop
Randy Bush

  - one held at RIPE 37, largely a tutorial function.
  - also one last week in Stockholm one of a series
    - purpose debugging software
    - working out sensible operational procedures etc Operational
      procedures seen as one of the very difficult parts of DNSSEC. 

Olaf Kolkman
    - RIPE NCC starting project to increase DNSEC awareness, develop  
guidelines etc.

Jim Reid (Nominum)

  - agreed with operational problems having a central place. 
  - also lots of extra NS work doing validation
  - answer packets much, much bigger due to signing etc
    -> dont fit in 512 bytes
    -> resolver gets truncated response
    -> repeats via TCP
    -> more server trouble.
    - waiting for EDNS0 to help with this
         -> larger packets
         - so far included in BIND9 only
  - parent must sign child zone key
    - how is he to be sure of the identity of requester,
      and that the request has not been intercepted and tampered with ?  
  - personal comments:
    - single failure point at root zone
    - also not much chance that signing/resigning will be done correctly and
      conscientiously when people can't even maintain simple delegations
      today.

5) RIPE NCC Report

Transfer of .de zone from ns.ripe.net to a server maintained
by DENIC using RIPE NCC AMSIX connectivity took place successfully.

6) CENTR Report

No CENTR representatives were present.

7) Anycast root server IP experiment
Masataka Ohta

  'Root Name Servers with Interdomain Anycast addresses'
  draft-ietf-dnsop-ohta-shared-root-server-00.txt
  draft-ietf-dnsop-ohta-shared-root-server-test-00.txt

Randy Bush

  - problem, when something goes wrong, where did it go wrong ?
    Which server had the bad answer
  - Masataka: traceroute
    - end users are anyway unlikely to attempt to troubleshoot
      in the event of failure anyway, but will consult their ISP.

8) Changing Domain Recommendations document
Fernando Garcia

  - a document to help with moving domains when changing IP addresses
    (e.g. changing provider)
  - helping guide, not standard or protocol.
  - draft sent to dnsop list Sept 10. Fernando wants to have final
    version ready for RIPE 38.
  - needs feedback to list, or to .
    The document can also be requested from the latter email
    address.

9) RIPE NCC Workplan

New workplan is out, document ripe-213.
   - new DNS item: DNSSEC activities.

10) DNS Software Status
Jim Reid

  - Nominum gearing up for BIND9 release
  - expects release probably by the end of 2000, perhaps sooner
  - BIND will remain open source.
  - BIND9 complete rewrite very little common code
   - threading important improvement. Previously NS didnt answer during load
   - also IPV6 support
   - differing internal/external copies of same domain on single
     server possible.
  - Plea: try the release candidates, havn't received much feedback.

11) DNS DOS Abuse

Nothing reported.

12) IO Other WG's or CENTR

Nothing reported.

13) Activity suggestions, suggestions for next meeting agenda ?

Nothing suggested.

Close