RIPE 43

Please mail comments/suggestions on:

• content to the Chair of the working group.
• format to webmaster@ripe.net.

Minutes (v1) ipv6 wg RIPE 43
Rhodes, Greece
11 September 2002


A. Administrative stuff
=======================

Agenda is available at:
http://www.kessens.com/~david/presentations/

The scribe was appointed, and the agenda was agreed upon.


B. Status of the 6bone (David Kessens (DK))
======================

Presentation is available at:
http://www.kessens.com/~david/presentations/

Per 20020911, there are now 1187 ipv6 sites in 58 different countries
on the ipv6 Internet.


C. IPv6 version of TTM (Henk Uijterwaal (HU))
======================

Presentation is available at:
http://www.ripe.net/ripe/meetings/archive/ripe-43/presentations/ripe43-ipv6-ttm/

Wilfried Woeber (WW): With 6net hat - is possible to make preliminary
measurements available before the end of the year?

DK: How is this service supported from the commercial point of view? Right
now, you have to pay for the test-box.
HU: You still have to buy a test box, and we operate it for you. Service
fee is currently 3000 EUR.

Q: Do you get v4 measurements with that, too?
A: In principle, yes.

WW: Technical question: the ntp protocol is used to encode v4 from clock-host
partners to prevent forming loops. Anyone here knows more?
HU: forward me the emails from the discussion.

DK: I like this new service. This can be a reason for me to order such
a box, because v6 production is important for us, and I want to be
able to measure the performance.


D. Action point 42.1:
=====================
Investigate the CNAME and other solutions for v6-reverse delegation

E. IPv6 capable RR DNS servers
==============================
What is needed, what are the issues ?!?
(RIPE NCC)

F. Action point 42.2:
=====================
Give an overview of 6-to-4 reverse delegation issues at RIPE 43
(RIPE NCC)

Andrei Robachevski, RIPE NCC covered topic D, E & F.

In regards to D & E, a prototype was made; setups how to support them
was tested; no final decision yet. Status: in progress.

It will be accomplished before the next ripe meeting

Mohsen Souissi (MS): Are you aware of the technical notes, published last
week, about DNAME, by the IETF committee? This is only meant for the
smooth transition from .int -> .arpa

Things are moving very very slowly. There is no dns server for ip6.arpa.
"ns.ripe.net" is not ipv6-capable.

I have already suggested several times to IANA that AFNIC server is
available to be a secondary for ip6.arpa, but I received no answer...

DK: There is an agenda point in the AOB about the Activity Plan of the
RIPE NCC. We need to add this in there. My suggestion is that the
RIPE NCC members make it clear to the RIPE NCC and the RIPE NCC
Board that we want RIPE NCC to give some real attention to this.


G. Global IPv6 routing table status (Gert Doering (GD))
===================================

Presentation is available at:
http://www.ripe.net/ripe/meetings/archive/ripe-43/presentations/ripe43-ipv6-table/

Q: Do you see people changing over, or announcing both?
A: It was possible to see everything, all combinations.

Q: Are the filtering problems still there?
A: Not really, but old prefix was still in the table, although I was
not announcing it any more...

Q: Do you see any difference between number of announcements vs number of
allocations per region?
A: This was not researched, it's a good idea for the next time.

Bernard Tuy (BT): While discussing the filters - we are at the point when
we do need some recommendations from the RIPE community.
DK: 6bone/IETF docs exist. Do we need new ones?
GD: They need re-doing, can not be accepted by default.
DK: Do we need recommendations?
GD: We need to at least keep an eye on this, and think about it. We
also need to think about multihoming issue. Let's put it on the
agenda for the next meeting.
DK: Any opinions right now?

GD: Just one customer is doing it - "announcing more specific
prefix" ?? I find it a good solution.

Francis Dupont (FD): That does not scale. It is too easy. ISPs do not
want to buy GB of memory to support this. But there's no good
alternative solution.

GD: Multihoming needs to be done by BGP : either /32, or /48 that can
get filtered. There are two groups busy with it in the IETF,
official&dead, and another one...

Proposal for the presentation on the next meeting - "Approaches to
multihoming and why everything sucks".

FD: "ipv6mh" is good place to discuss this. It's a closed mailing list.

DK: Personally, I think the way Gert is doing it is a good way.

DK: Anyone who is a good candidate to lead the discussion, present
existing solutions and explain why they don't work, please
volunteer for the next meeting.


H. Report(s)
============

About *actual* v6 traffic volume as compared to v4?
What's *real* out there, not what's on power-point?

H.1 baseline measurement for ipv6.int queries
----------------------------------------------
(Terry Manderson, APNIC)

[URL - Presentation missing!]

Q: How up to date are your graphs, and can you see RIPE43 meeting
influence?
A: From last Friday, so can't see this meeting.

Q: different considerations - DNS traffic as protocol, and v6 traffic
after DNS queries. how to get v6 related RRs. unless you have v6
from the root, except the glue, you don't know where else to go.

Q: Can you see where are the queries coming from?
A: These were the baseline measurements, not analysis. But, most cam from
Japan.

MS:.jp & .fr have been supporting v6 in DNS for years. We are asking
for glue records in root, but according to the policy that is not
possible. We are again asking RIPE NCC to support our efforts to
tell to the root - v6 is harmless.

DK:In the Activity Plan input to the RIPE NCC, we should be making it
clear that this is important to us, as members of the RIPE
community.


[coffee break]


H.2 some snap-shot info
----------------------
(Wilfried Woeber)

[URL - Presentation missing!]

DK: For the next RIPE meeting we will be preparing dns/v6/sec workshop.


H.3 Discussion
--------------

Do we want some more activity under this topic - to cover it in depth; to
have more data? e.g TTM with more data.

Since this is a _working_ group, more work from the participants is
appreciated...


I. Developments/initiatives
===========================
regarding IPv6 in the RIPE region and beyond
(input from the audience)

I.1 Bernard Tuy, Renater
------------------------

We are deploying M6bone. 40 players, 15 non-fr countries. If you are
interested, you are more then welcome to join us.

http://sem2.renater.fr/m6bone/
What do we want to achieve, what kind of SW is used...


I.2 Robert Blechinger, CyberNet
-------------------------------

Number of queries per day, on their DNS server: 28 6bone, 2 6to4, 336
RIPE NCC allocated: 211 own, 115 from one peer.

I.3 Mohsen Souissi
Ready to publish DNS extensions to support IPv6 (updating and obsoleting RFC 1886) to draft standard status.

I.4. DK
The same week that there is a RIPE meeting in Barcelona in May
there is a IPv6 Forum in Madrid. We are trying to schedule that at
least IPv6-wg does not overlap.

MS: This might be an opportunity to test video-conferencing over IPv6.


Z. AOB
======

Z.1. input to the RIPE NCC Activity Plan
-----------------------------------------

Please let your representatives tell to the RIPE NCC management and
Executive Board your requirements, so that they can be included in the
Activiy Plan.


Z.2. 6bone transition to the RIRs
----------------------------------

There has been a talk in the LIR-WG regarding that. 6bone addresses are
currently allocated by Bob Fink, which is inconvenient as being a single
point of failure, and is outside of the RIR structure. If interested, make
sure to take part in the discussion on the lir-wg list & meetings. The
progres of the co-ordination efforts will be reported on the next meeting.

The most important are the technical issues - ip6.int & ip6.arpa delegations.
They will be put on the agenda for the next meeting.


Z.3. Other
----------

Info: Bernard Tuy:
29 & 30 October, IPv6 conference in Paris
28 October, IPv6 Tutorial by Mohsen Souissi, like the one done on Monday
in Rhodes.

Andrew Miehs, CyberNEt: ipv6-wg meeting should not be at the same time
EIX!! One of the issues - we'd like to establish peering points, but I
can't express this in the IEX-WG.

DK: Other people asked for this too, many many times.

The same - DNS-WG also has many v6 topics now, and is also scheduled at
the same time!

WW: Instead of complaining, it's better to put a proposal - how can we
improve the scheduling?

Action point on DK: to bring the issue to the WG chair lunch, and try
to find a solution.