Plenary Session Minutes

Plenary Session Minutes from RIPE 39 which took place in Bologna.

Chair: Rob Blokzijl
Scribe: Sabine Mader

1. Opening and Welcome
2. Agenda
3. Minutes of RIPE 38
4. From the Chair
5. Operational Report from the RIPE NCC (Axel Pawlik)
6. Reports from other Regional Internet Registries
7. Global RIR Statistics
8. Reports from ICANN and the ASO
9. Routing Observations
(Geoff Huston, presented by Philip Smith)
10. The Commoditisation of the Telecommunication Industry
11. Reports from the Working Groups
12. European CERT (Wilfried Woeber)
13. Next Meetings
- RIPE 40, Prague, 1-5 October 2001
- RIPE 41, Amsterdam, 14 - 18 January 2002
14. AOB
15. Closing

1. Opening and Welcome
Rob Blokzijl welcomed the attendees to the RIPE 39 Meeting.

2. Agenda
The RIPE 39 Meeting participants approved the agenda.

3. Minutes of RIPE 38 Meeting.
The minutes were approved.

4. From the Chair
The Chair gave an overview of the activities during the RIPE meeting so far and expressed his thanks for the successful organisation of the meeting. Special thanks from the RIPE NCC staff to the best network host they ever had,
which managed in a tremendous way to deal with the network outage caused by a heavy thunderstorm.

The Chair relates to a current discussion on the RIPE-list regarding the RIPE Document Store. The RIPE NCC made changes in some of the documents and published them without changing the number for these documents. This was an error and has been corrected.

The rules for the RIPE Document Store remain: once a document is issued, it will not be changed ever again, not even spelling mistakes. This makes it a bit more complicated to refer to documents that change frequently. It has been decided that we will not introduce different versions under the same ripe-doc number.

However, we want to introduce a system that makes it easier to refer to document titles (specially those that change more frequently). Rob realises that there was an action on himself to work with the RIPE NCC on a better system which he has not finalised, but will do this very soon.


5. Operational Report from the RIPE NCC (Axel Pawlik, Managing Director)


There were no questions.

The Chair expressed his thanks to all RIPE NCC staff for their work.


6. Reports from other Regional Internet Registries

- Ray Plzak, ARIN

Questions:
Does ARIN or the ARIN community care about getting TLD DNS servers or IPv6 capable root servers?

Ray: No, ARIN is not involved in the operation of the ccTLDs in its service region. The RIRs generally are not involved in the forward DNS, but we know that from the perspective of ARIN, and together with the IETF and the other registries, they do start working at the IPv6 reverse delegation.


- Anne Lord, APNIC
Anne Lord gave a progress report update on APNIC.
There were no questions.


7. Global RIR Statistics - Nurani Nimpuno, RIPE NCC

Nurani Nimpuno gave a combined RIR report on their statistics projects.

All statistics are available on the RIPE NCC website.

Questions:
Wilfried Woeber: Regarding pie chart of other organisations (3 pieces RIPE NCC and APNIC), is the legacy address space part of other organisations or is it part of the
ARIN (second of the pie) ?
Nurani: The second are the different RIRs.
Wilfried: The legacy part is in the other part of the pie?
Nurani: That is correct.

8. Reports from ICANN and the ASO

- Stuart Lynn, ICANN
The Chair expressed his pleasure to introduce Stuart Lynn, the new president and CEO of ICANN, to give an update of ICANN.

Stuart Lynn gave his first official presentation since he took over as ICANN president.
There were no questions.

- Mirjam Kuehne, RIPE NCC (ASO Secretariat)
Mirjam Kuehne gave an overview of the ASO activities.

There were no questions.

9. Routing Observations
(Geoff Huston, presented by Philip Smith)

References: http://www.telstra.net/gih/papers/ietf50-bgp.pdf

Questions:
Bill Manning: Dynamic nature of routing system is killing
Sean Donelan: A lot of the long prefixes in space that could be aggregated.

Business model: sending letters to those organisations. Threatening them to charge them if they do not fix this, because it costs the transit ISPs money. Suggests that RIPE or the RIPE NCC produce a document on why it is
good to aggregate.

10. The Commoditisation of the Telecommunication Industry (Daniele Bovio)

Questions:
Bill Manning: I'm encouraged to speak about California. About 15 years ago, there was a statement made that there was so much electricity that it would be too cheap to deal it. There were some fundamental assumptions  in what you stated. Bandwidth is not possible for me – as a resident of the state of California – to decide that I would like to dig fibre in Bologna or wherever I want. Politics and all the present things that dictate where you could put transmission
capacity. That is not an open market as people would really believe.

Daniel Karrenberg: Daniele mentioned that bandwidth could be sold on a weekly or monthly basis. Why not on an hourly basis? Maybe I only need bandwidth for a conference call for tomorrow?

Daniel: Yes, no problem. Telephone companies are used to make 3 or more years contracts. Therefore, it may mean quite a change for them to even make contracts on a monthly basis.

Rob Blokzijl: That might be a specialised sub-market to sell only hours.

Sean Donelan: Bandwidth is not comparative, there are no standards. It is difficult to build commodity on something that does not physically exist. Also, there are no standard bandwidth exchanges.

Daniele: I agree with this concern, this is not something that will happen overnight. Message is: it looks like that there is a serious chance that this will happen, maybe even in the near future.

Harald: Do you expect that these exchanges will happen on layer 2 or on layer 3?

Daniele: I believe on both layers.

Bill: How many pooling points will there be?

Daniele: Depends on the demand and requirements on the market. It does not take much to set up a pooling point.

Bill: A number of pooling points exist in residence with layer 3 exchanges. There are a number in the traditional copper arena which one can call layer 3. Co-location space is not enough. Only providing space is not sufficient. You got to have neutral space in attractive locations.

Bill: Pair is too fine grained. Bits between major exchanges is better.

Daniel Karrenberg: How about between ASes?

Bill W: AS borders yes, an AS border pair is essentially saying between the two.

Sean: Has customers that sit at IXPs. No problems building a contract: use this line until you hit 1 million packets. Would be harder to do it more fine grained. Quite workable to set up anonymous IXPs.

11. Reports from the Working Groups


The Chair reminds the WG Chairs to provide the RIPE NCC with a short summary in order to include them in the minutes.

After an observation of Working Group agendas and several discussions with WGs during this week, he stated that an increasing number of WGs is dealing with security-related issues. He invites all WG Chairs and participants to start
thinking about this and to come with ideas between now and the next RIPE meeting on how we can more efficiently deal with various network security aspects.

He expressed his hope that WG Chairs will include some commentary conclusions into their reports. He intends to put this topic on the RIPE mailing list, as well as to invite the whole RIPE community to give us some input on how we might co-ordinate our various security-related efforts better in the near future. It will also be discussed during the planning of the agendas for the next RIPE Meeting.
These security issues will then be discussed further at RIPE 40 in Prague.

a) DISI BoF (Daniel Karrenberg)

A new mailing list has been set up and announced for the DISI (Deployment of Internet Security Infrastructure) project. Further developments and the idea to set up a new WG will be discussed on the RIPE mailing list.

There were no questions.

b) EIX-WG
There were no questions.

c) NetNews-WG (Dave Wilson)

-Originally not planned to meet this time. However, it turned out at the RIPE Meeting that there was interest after all.

Questions:
John Brown: How much news is actually useful (signal-noise ratio)? Can we make it more useful to restrict abuse of the protocol?

Dave: Depends on the definition of what is useful and what is abuse.

John: It might be an idea to put a large news server environment at IXPs so that the members of the IXPs can share this. This has quite successfully been tried at some rural IXPs in the US.

d) DB-WG

Wilfried encouraged everyone affected to discuss these procedures together with the RIPE NCC on the mailing list.

There were no questions.

e) LIR-WG (Hans Petter Holen)

f) Tools-WG (Joao Luis Silva Damas)
There were no questions.

g) Routing-WG (Joachim Schmitz)
There were no questions.

h) IPv6-WG (David Kessens)


Questions:
Paul Mylotte: There are two solutions for assigning address space for IXPs. Will we keep both?

David: No decision has been made about this.

i) DNR Forum (Fay Howard)
There were no questions.

j) Test Traffic Measurements-WG (Olaf Kolkman)

There were no questions.

 

k) Anti-Spam (Rodney Tillotson)
There were no questions.

l) DNS-WG (Peter Koch)
There were no questions.

12. European CERT (Wilfried Woeber for Don Stikvoort, CEO Stelvio)

Wilfried Woeber gave an overview update on the European CERT activities.
There were no questions.

13. Next Meetings
Next RIPE Meetings will be held as follows:
- RIPE 40, Prague, 1-5 October 2001
- RIPE 41, Amsterdam, 14-18 January 2002

14. AOB
The Chair expressed his special thanks to Compaq, Cisco Systems and Telecom Italia for the excellent terminal facilities. Sincere thanks were also expressed to INFN, Level 3, Acamai, and Ebone for their assistance in organising all social events and for their sponsorship.

The local installation of the network systems was greatly done by Gabriela Paulini, Andrea Sarpati and the RIPE NCC.

15. Meeting Close