RIPE 44

RIPE Meeting: 44
Working Group: IPv6
Status: Final
Revision Number: 2

Minutes IPv6 working RIPE 44

Meeting: IPv6 Working Group - RIPE 44
Date: Wednesday, 29 January, 2003
Time: 2pm - 4pm
Chair: David Kessens (DK)
Scribe: Timothy Lowe

Webcast: rtsp://webcast.ripe.net:7070/Archive/ipv6-1.rm

A. Administrative stuff
- Agenda bashing
Agenda available at: http://www.kessens.com/~david/presentations/
- No new agenda updates were proposed

B. IPv6 version of TTM
- Henk Uijterwaal

http://www.ripe.net/ripe/meetings/archive/ripe-44/presentations/ripe44-ipv6-ttm/

- No questions or comments were made

C. Presentation - IPv6 whois proxy
- Shane Kerr (SK)

http://www.ripe.net/ripe/meetings/archive/ripe-44/presentations/ripe44-ipv6-v6wpd/

Gert Doering (GD):
Will IPv6 be integrated into the main whois server, not
just on the proxy server?

SK: Those are our future plans but their some issues including
security issues that need to be worked out first.

DK: Don't put all your efforts into proxies.

SK: We won't.

D. IPv6 state of the art in the Italian educational
and research community.
- Gabriella Paolini (GP)

Francis Dupont:
Don't use /126 for point to point links

GP: It avoids unnecessary waste.

Speaker: (Did not use microphone)
It includes loopback and network so it is

DK: We can come up with a best current practice document on numbering
v6 networks as an agenda point for next time.

E. Global IPv6 routing table status
- Gert Dvring (GD)

Bernard Tuy (BT):
Improving routing is something we should be working on.

GD: Yes maybe a best current practice document.

BT: Do we want a guideline document?

GD: It would be useful

DK: There is a 6bone document that describes this.

BT: I want a RIPE document on this and volunteer to write one

DK: Action on Bernard to co-ordinate for a routing document.

Philip Smith (PFS):
Would an IPv6 CIDR Report be useful?

DK: Yes

PFS: Fine, then Geoff and I will get on that.

PFS: We provide transit to anyone at all right now. We want to
change this we can discuss this later.

DK: I am also encouraging people to look for other transit then
from me.

GD: I wish to keep the link to Cisco now with the shorter path
lengths.

PFS: All right I would appreciate any feedback.

F. Ghost Route Busters
- Jeroen Masser (JM)

ttp://www.ripe.net/ripe/meetings/archive/ripe-44/presentations/ripe44-ipv6-ghosts/

Randy Bush (RB):
What is TLA?

JM: Top Level Aggregater

RB: What?

A short exchange on the deprecation the TLA syntax followed.

G. Traffic graphs
Jeroen Masser

Presentation was not given but is available from:

http://www.ipv6.aorta.net

H. 5 year trend in the BGP table
Ronald van der Pol

http://www.nlnetlabs.nl/ipv6/measurements/index.en.html

I. Workable approaches for IPv6 multihoming in the absence of
a better solution than the one we use in the IPv4 Internet
- Iljitsch van Beijnum (IvB)

http://www.bgpexpert.com/ripe44/

BT: Why is the draft not out?

IvB: The IETF can't agree on it.

BT: We could write a draft here?

DK: Bernard that is IETF work

IvB: It is possible that we could submit a draft to them but
let me finish here my presentation and we can discuss it
later.

Clemens Zauner (CZ):
Putting multi addressing and multihoming in the same
presentation is not so logical.

IvB: Are you saying if you are doing multi addressing you are
not really doing multihoming?

DK: There are people that say if you do multi addressing right now you
don't need to multihome.

CZ: Also most of the deaggregation taking place is not due to
multihoming.

IvB: I agree this premise is false even if every ISP announces
10000 routes ....

CZ: Even so I don't belive in millions of multihomers.

IvB: Neither do I but v4 style multihoming won't scale.

CZ: Even so I don't belive multihoming will break things if you
look at it now.

IvB: Yes but we also have to think 20 years ahead.

CZ: Maybe we could make them prove a need for multi homing

IvB: There is the possibility of coming up with an administrative
solution like that but one solution could be to make them pay
a fee. You cannot have an upstream for free whether it is
one or many upstreams.

GD: There are a number of routes that shouldn't be there but are
for traffic engineering reasons. A reason for multihoming that
I see is that they want connectivity redundancy due to ISP
bankruptcies.

IvB: I want to take this discussion offline to go into it further

J. IPv6 Multihoming part 2
- Kurt Erik Lindqvist (KEL)

http://www.ripe.net/ripe/meetings/archive/ripe-44/presentations/ripe44-ipv6-multihoming/

GD: I like this idea. of increasing the allocation from IANA to
the RIR's. Make it an action on the RIPE NCC to talk to IANA
about this. If this gets out of hand we can roll it back. It
is pretty radical.

KEL: It is not a long term solution but it buys us time.

IvB: 64,000 multihomers is not the upper limit.

KEL: Even so it still won't break routers at twice as many.

DK: Should we go to routing with this?

GD: Yes.

KEL: Yes

K. Discussion point - Fragmentation of the IPv6 address space from
the top down by allocating /23's to RIRs
- Gert Dvring

AOB

DK: Another IPv6 survey is in the works and the RIPE NCC needs
input on what questions to ask in the survey. Please give
them input.

BT: Presenting useful data is a good idea if you have any ideas on
this pass them on. Also we have a v6 project the M6bone please
come play with us.

DK: We arranged this working group session to avoid overlap with
the other working group sessions and we have moved topics to
other appropriate wg's. We would like input on if you like this
approach. Please inform me.

I. Workable approaches for IPv6 multihoming in the absence of a
better solution than the one we use in the IPv4 Internet
- Kurt Erik Lindqvist, Iljitsch van Beijnum

brief discussion

K. Developments/initiatives regarding IPv6 in the RIPE region and
beyond input from the audience

L. Input to the RIPE NCC Activity Plan & results of KPMG survey
input from the audience

DK: It is very important that the RIPE community tells the RIPE NCC
what IPv6 activities should be put into the activity plan

Z. AOB

Agenda point for RIPE 45 best current practice document on numbering IPv6
networks


Action Points from Previous meetings:

42.1: Investigate the CNAME and other solutions for
IPv6 reverse delegation - RIPE NCC
42.2: Give an overview of 6-to-4 reverse delegation
issues at RIPE 43 - RIPE NCC
42.3: IPv6 capable RR DNS servers - what is needed,
what are the issues? - RIPE NCC

Action points 42.1-3 were rewritten and closed.

New Actions Points:

44.1 Enable reverse delegation For 6bone (RIPE NCC)
44.2 To co-ordinate an IPv6 routing guideline document (Bernard Tuy)
44.3 To investigate with IANA increasing the size of RIR allocations to
rationalize routing (RIPE NCC)
44.4 Report back to the working group what the status is and what the issues
are regarding:

- enabling AAAA glue records in the . zone
- AAAA glue records for delegations to TLDs
- AAAA glue records for the root servers themselves
- enabling ipv6 transport to the . zone

and come back with ideas on how these issues can be overcome (RIPE NCC).