- content to the Chair of the working group.
- format to webmaster _at_ ripe _dot_ net.
IPv6 Working Group Minutes, RIPE59
Date: Tuesday 6-10-2009, 16:00-18:00
Chair: David Kessens
Minutes: Alex Band, RIPE NCC
Jabber monitor: Rumy Kanis, RIPE NCC
David welcomed the attendees and opened the session.
A. Administrative Matters
• Select scribe: Alex Band
• Select Jabber Monitor: Rumy Kanis
• Finalise agenda
• Approval of minutes from previous working group meeting
B. RIPE NCC IPv6 Update
James Aldridge, RIPE NCC
There were no questions.
C. Future of the Working Group: Charter Discussion
Shane Kerr, David Kessens. Input from the audience
Shane mentioned that the initial idea for an IPv6 Working Group came from looking at the statistics Geoff Huston publishes. "Make IPv6 Happen", was the summary of the charter of the WG. Most of the discussion on IPv6 is actually in the Address Policy WG, so he proposed to shut down the WG and move on to bigger and better thing.
David asked the audience for input.
Olaf Kolkman (NLNetlabs) asked how the activities of the WG tied into the IPv6ActNow website.
Shane said that there is not really any tie in, and that he thought that is was a shame that something like IPv6ActNow is necessary at all.
Rob Blokzijl (RIPE Chair) mentioned that the message that is coming out of the WG is heard by the community at large, however, it doesn't seem to be received and digested by the industry on a wide scale. He proposed that the WG is not shut down, because there is important work to be done, especially in the coming two years and when IPv4 has run out.
Gert Döring (Spacenet) commented that the WG is not the place where people go nowadays to learn about IPv6. At the RIPE Meeting, those kinds of discussions are held in the Plenery.
Rob gave an example: there were two presentations about the general implementations of IPv6 at customer premises. In one case, the CPE worked fine, in the other it didn't. He would have liked to have seen a technical presentation in the IPv6 WG session on why that is.
Shane responded there are definitely opportunities for the WG in that capacity, and even other topics related to IPv6, like Multi-layer NAT.
Denesh Babutha (Aexiomus//Cyberstrider) agreed that the WG had reached the end of its useful life as originally envisaged. He said that, as such, the WG needs to evolve, and he thought that this needed to happen through the the charter. Denesh added that the WG needed to move away from repeating the IPv6 situation in every single meeting and move onto more useful discussions. The new charter should be inclusive of other industry areas too.
David said that there were many topics covered by the WG that were not originally in the charter.
Bernard Tuy said that there is an enormous need for technical training and that kind of knowledge can come from the WG.
David asked the audience how to move forward with the WG and whether it should continue in it's current form?
Ruediger Volk (Deutsche Telekom) commented that the current form seems to be "quasi-plenery".
Rob commented that the WG should be modest in its aims and not try to reach the end goal without some intermediate steps. He said that the WG should focus on real world implementations.
Shane commented that the WG did not need to come up with a new charter here and now.
Rob proposed that a new draft charter is written.
David summarised the feedback for a new charter as more focus on education and outreach activities.
Maarten Botterman said that not all of those activities have to be within the WG.
David asked the audience if preparing for a hybrid IPv4/IPv6 world should be part of the charter.
Rob commented that outreach should lie with the RIPE NCC efforts, such as the IPv6ActNow initiative. He said that training is a sensitive issue because the RIPE NCC does not want to compete with its members.
Rumy Kanis (RIPE NCC Training Services Manager) commented that the RIPE NCC offers IPv6 training, but careful steps were taken to ensure the RIPE NCC does not compete with the membership.
Marco Wertejuk (Binconsult) said that he believes the focus for the new charter should be "how to encourage conservation of IPv4 and deployment of IPv6".
David said that he assessed from the audience feedback that there isn't much support for the WG to focus on non-technical outreach activities. David asked Rob for a timeline.
Rob said that he thought it was important that the people who could not attend the meeting could have a chance to comment on this. He added that this could be discussed on the mailing list over the coming months and put in place before the next RIPE Meeting so that the next working group session could take place under a new charter.
D. Future RIPE Network Experiments - David Kessens. Input from the audience
David asked the audience if there was further interest in IPv6 network experiments during a RIPE Meeting.
Gert said that forcing IPv6 on people is actually a good idea. He said we should even go as far as making the RIPE Meeting network IPv6 only for the whole week. He added that we should do the experiment without NAT-PT.
David said that he would like to see a completely new experiment.
Marco Hogewoning (XS4all) said that although IPv6 connectivity in general is good, tunneling is needed in some places in order to get full
David Wilson (HEAnet) said that the IPv6 hour should be repeated, and that the WG should get new experiences because things have changed since the last time it was done.
Izumi from JPNIC that for their experiments, they created different testbeds for different kinds of implementations.
David asked the audience if the IPv6 Hour should be rerun at future meetings. There is consensus that it should be.
E. IPv6 Deployment: What are the Remaining Issues and Bottlenecks? - A panel discussion with interactive input from the audience.
Moderator: Maarten Botterman, GNKS Consult
Gert Döring introduced himself and asked the question "What kind of Internet do we want to have in five years?".
Geoff Huston introduced himself and in a slide set proposed the question "Is the IPv6 transition an example of market failure?"
Kurtis Lindqvist introduced himself and explained his own experiences with IPv6 deployment. He said that doing a slow and gradual implementation has been very beneficial.
Maarten Botterman did a live survey, with the question "What would you gain by postponing, and what would you gain by stepping up to the plate, earlier rather than later?". 40 votes were cast, most go to:
• By stepping up to the plate I have time and space to do the transition, which will allow me to do it gradually and well thought through
• By stepping up to the plate I confirm my brand’s image as “state-of-the-art” or “top-of-the-wave”
Geoff commented that the respondents are most likely non-representative.
Gert said that he is not surprised with the numbers.
Kurtis said that looking at the responses, he is optimistic.
Maarten said that some people responded with 'none of the above'. He
asked three of those people to comment:
Ruediger said that for all of the relevant players there is still enough time for doing a proper job.
Bernard said that there was no vendor support, so ISPs will never start until that happens.
Jan Hugo Prins said that there will never be end user demand. Deployment is not picking up because of security issues.
Geoff reiterated that he thought that the situation we are finding ourselves in right now is no accident. It is in many companies' interests to delay IPv6 adoption.
Shane agreed with Geoff.
Gabriella Paolini (GARR) said that the biggest issue is that middleware doesn't support IPv6 and that changes very slowly.
Maarten did another live survey with the following question: "There is perception and experience of/with a number of hurdles that seem to be difficult to overcome. Which are important to you?".
• As there is very little customer demand, I cannot justify the investment to my management, even if I would like to move
• I have the experience that my vendors cannot deliver what I need
• None of the above (highest score)
Ruediger said that the options are very extreme. Slow and gradual deployment is a real option.
Geoff commented that in his days at a telco, the biggest cost turned out to be customer support. IPv6 could become expensive in that respect with end users calling in because of issues.
Maarten asked people to comment:
- Wilfried Woeber (Vienna University) commented that he would have to select 'none of the above' for all questions. His issue is that the questions are too black or white. The biggest problem from his point of view is in layer 7 or 8: infrastructure, applications etc. The backbone has been done for years.
- Marco said that he agreed with Geoff, the cost is not in the hardware. IPv6 is not a product, don't pay extra for it.
- An attendee said that in a lot of cases, even though the product spec sheet says that IPv6 is supported, there is no feature parity with IPv4.
Gert said he agreed with most of the comments.
Geoff said that any vendor is willing to give you what you want, as long as you put down the money.
Kurtis added that there was also a big push from vendors for MPLS.
Marco added that some IPv6 functionality used to be free, but after vendors realised that there was a market for it, they started charging for a license.
Maarten did another live survey, the results with the most votes were:
• The deployment of IPv6 is unavoidable for my organisation in due time
• Technically, everything is in place to deploy IPv6
Geoff commented that openness is the key in the end. IPv6 is the only way to guarantee that.
Kurtis added that he was surprised that so many people responded "I will deploy in due time", because that is pretty close.
Geoff re-stated that the market is well informed, and that this situation is intentional in order to maximize shareholder profits at large corporations.
Ruediger agreed that the industry is flooded with information. But he asked how enlightened are people?
Geoff responded that it's all about vendor lock in.
Brian Nisbet (HEAnet) said that the WG keeps having the same discussion meeting after meeting. He said that he could not believe that people are not well informed.
Kurtis said that CFO's cannot be convinced of the urgency of something that will happen two years from now. An attendee said that for mobile operators, the vendors of mobile terminals are the blockers.
Maarten showed the results of the last survey. The panel shared some closing thoughts. The general thought was that the community should stop distrubuting messaging, and start moving packets.
Maarten closed the discussion.
F. Developments/Initiatives Regarding IPv6 in the RIPE Region and Beyond
- Input from the audience
No input was given.
Y. Input for the RIPE NCC Activity Plan - Input from the audience
No input was given.
David closed the session.