[ipv6-wg at ripe.net] Draft Minutes (v1) IPv6 WG RIPE 50

Please see below for our first cut of the minutes for our working
group session during RIPE 50.

Please let me know if you have any comments or corrections.

The minutes will be declared final by Friday, June 10, 2005 if no
significant changes (other than typo fixes and minor editorial issues)
are found.

David Kessens
---

Draft Minutes (v1) IPv6 WG RIPE 50

When:         Friday, 6 May 2005, 09:00 - 10:30
Where:        C58, Clarion Hotel, Stockholm, SE
Minute Taker: Adrian Bedford


Quick Update from the RIPE NCC Regarding IPv6 Services
(Ruud de Kooter - RIPE NCC)

After the presentation, Ruud was asked if there were plans to
integrate the IPv6 whois interface with the main whois. Currently the
proxy cannot talk to the interface and this hampers some functionality
(such as rate limiting). Ruud asked Shane Kerr to answer this. Shane
explained that the RIPE NCC has a v6 proxy running over a v4 database
backend. The RIPE NCC has been looking at how people use the database
for two years and patterns suggest that now would be a good time to
integrate native IPv6 into the server using /48 as a client identifier.


Discussion on: Global IPv6 Routing Table Status
(Gert Doering, Spacenet)

Gert indicated that the Routing Working Group was discussing best practices
for BGP filters. The overall thrust of the presentation was that v6 is
growing smoothly with no big disasters so far. A question was asked about
ratio comparison between the numbers of Autonomous Systems (AS) announced
by IPv6 networks as opposed to v4. The current data suggests that there are
1.419 announced entries per originating AS for v6 and 10.308 for v4.
A point was made that Deutsche Telecom (DT), who have one of the largest v6
allocations made to date would not announce their block as one AS, they
would be more likely to split this and allocate blocks to different service
providers for political reasons. Someone from DT commented that the company
has the right to do this, though denied that the move was in any way
politically motivated.


Renumbering IPv6 Networks
(Stig Venaas, Uninett & Tim Chown, University of Southampton)

After the presentation, it was mentioned that we are assuming that
customers of service providers will keep their space, yet ten years
ago major providers did force renumbering onto customers. It was
agreed that renumbering of IPv6 is, at least in principal, no
different to what happened with v4.


Developments/Initiatives Regarding IPv6 in the RIPE Region and Beyond

- Francis Dupont:
  Point6 (http://www.point6.net/
- Gert Doering:
  A new maillist for discussion of ipv6 operational issues was started:
  http://lists.cluenet.de/mailman/listinfo/ipv6-ops/ 

  After the announcement of the list, it was agreed that having a
  publicly available archive of this new list would be a good idea.
  Although it exists on the site that hosts the list, this is a
  personal website and for stability purposes, maybe a copy of the
  archive should be placed elsewhere. It was felt that the RIPE
  website was unsuitable for this. 

- Carlos Friacas:
  o IPv6 - Advanced Network Developments's first year in Portugal
  o 9K IPv6 Schools' Network Initiative
  o About the IPv6 Steering Committee


Revisiting the /48 recommendation

There was a lengthy discussion around RFC 3177. It was made clear that
the discussion was not intended to reach final consensus on all the
issues involved, but merely to serve as a way to get an idea about the
general direction on where the community would like to go to make it
possible to write up an initial proposal.

Many agreed that the /48 is too much for many users for now and in the
future. A discussion followed about how we should define a site with
the suggestion that we need to accept that this can remain a grey
area, which is not at all well defined right now. Discussion is needed
around this topic, though the group tended to accept that it might not
be easy to put down limits in black and white.

Most people in the room felt that RFC3177 needs a revision. Most
people felt that the /48 limit itself doesn't need to be changed, but
most agreed that there is a need for a new category that falls
somewhere between /48 and /64 for users that have a need for
subnetting but for which a /48 seems excessive. This category would
fit the mass consumermarket or (very) small office market, however, it
was noted that we should base the category on technical needs of the
user, not in economic terms like mass consumer market.

It is exepected that an Internet Draft will be written that can be
discussed at RIPE 51.