IPv6 Working Group Minutes RIPE 71

Session I

18 November, 14:00 - 15:30
Scribe: Alex Band
WG chairs: Jen Linkova, Benedikt Stockebrand, Dave Wilson

A. Welcome and Administravia

Jen Linkova opens the session.

B. The Next Step: IPv6 from the Enterprise Perspective - Benedikt Stockebrand

The slides are available at:
https://ripe71.ripe.net/presentations/119-2015-11-18-RIPE71-IPv6-at-Enterprise.pdf

Jen Linkova asks if it would be useful to have a page with Frequently Asked Questions (FAQs). Benedikt replies that usually, the people who read an FAQ do not need convincing. All you can really do is spread the word. Most people will only realise they have to do something with IPv6 once it starts to hurt their business. Raymond Jetten from Elisa says there is information available for management on why they should do IPV6. The biggest problem is convincing organisations where all the money is reserved for making the shareholders happy. Benedikt says this is true and even worse, there is a lot of short term thinking.

C. IPv6 DHCP for ISPs - Sander Steffann

The slides are available at:
https://ripe71.ripe.net/presentations/73-DHCPKit-RIPE-IPv6-WG.pdf

An audience speaker says that he is interested in the work that Sander presented. They are currently using separate IPv4 and IPv6 DCHCP servers, which they know will not be reliable for the long term. Considering Sanders work, they are especially interested in the RADIUS part. Piotr Strzyzewski from Silesian University of Technology, Computer Centre would like refers to the fact that Sander said that the existing DHCPv6-PD weren't good enough. Piotr would like to know which DCHP server were tested and if he tried KIA. Sander says we looked at ISC, Dibbler and WIDE but not KIA. He continues that KIA would work in this setting. The biggest problem is that all the tools work with leases files, which can cause problems when swapping CPEs; it refuses to give a prefix to the new CPE because it still has a lease for the old one. Piotr finished by saying that in terms of the landscape, KIA is only competition right now. Sander agrees with this comment.

D. Lightning Talks

D1. Flow-based Detection of IPv6-specific Threats - Luuk Hendriks

The slides are available at:
https://ripe71.ripe.net/presentations/104-flow-based-v6-specific-threats.pdf

Erik Bais from A2BInternet asked if Luuk also had a look at the IPv6 headers in the packet as some switches were limited by the ACLs. He wants to know if people are using the four standard headers and then add additional ones to get the information they actually want. Luuks says that they haven't looked at that yet, but you can do so much nasty stuff with the headers it should be a lot of fun. The fragmentation aspects of the research was a first step in that direction, e.g. if it's protocol 44 then they trigger their logic. Jen Linkova adds that many devices are not capable of looking at additional headers after the first one. If you add a lot of headers, you actually run a good chance of your packets being dropped.

D2. IPv6 BCOP for Enterprises - Sander Steffann

Rob Evans from JANET comments that there is a document on IPv6 route aggregation (ripe-532). Sander says they'll definitely use it. Jan Zorz from ISOC has been working on a similar project and have been sending draft documents to the community asking for feedback. Jan is asking what the best format is for requesting feedback. Sander suggests to start with the BCOP task force mailing list to discuss the content. Sander also wants to add the option for people to leave notes on the website. Jen Linkova adds that they had some very positive experiences working on help desk documents with regards to BCOPs. In addition, she suggests that the IPv6 Working Group is not excluded from the discussion. Sander guarantees that won't happen. Jen would also like to know if Sander is considering Github for version control. Sander says this is a good idea. Benedikt Stockebrand mentions that he used video blogs, he wonders if that is any use to Sander. Sander says that all content that is useful for enterprises is welcome.

D3. IPv6-only BCOP - Jen Linkova

The slides are available at:
https://ripe71.ripe.net/presentations/74-IPv6-only-Eyeballs-Networks-BCOP.pdf

E. Open Mic Discussion: IPv6-only Network @ RIPE Meetings: What's Next? [10 min]

The slides are available at:
https://ripe71.ripe.net/presentations/103-RIPEMTG-NAT64.pdf

Marco Hogewoning mentions the network needs to be perfect operationally, meaning that for example all VPN clients that currently don't work on the IPv6-only network pose a real problem that needs to be resolved. This is something that needs to be looked at on the longer terms as well, for example wit regards to hardware replacements over the coming years. Jen is happy to see it is actually possible to offer an IPv6-only network without the sky falling. Jan Zorz from ISOC would really like to see a detailed report on what issues there currently are, so we can make a plan to resolve them. Jen says she is currently collecting this information and putting it on a Wiki page. Jan adds that this Wiki information is also interesting for people outside this community. Marco says that he is especially concerned about the VPN issues. Jen says some issues can be resolved easily, other are more complex. Sander Steffann apologised for being blunt on the mailing list and applauds the work that is being done. Benedikt Stockebrand says that the IPv6 network has been established 15 years ago to push it to the extreme. Delaying things doesn't make it any better in his opinion. Then again, every RIPE Meeting a little step forward is taken, which is great and should not be stalled. Benedikt expects this discussion to continue for a couple more years. In the end, don't take criticism personally. Hen would like to know if the community can do anything to help the RIPE NCC to help getting the job done. Marco says that it's especially important to also indicate when a certain issue has been resolved, so not just issues, but also solutions. Jen also wonders why it was so explicitly mentioned that no support would be provided on the IPv6-only network, because it might scare people off. An audience speaker adds that some feedback works better when it can be provided in physical form, like a card. Marco says there were flyers. Shane Kerr from BII says it should be feasible to increase the numer of IPv6 users by the next meeting to 20%. He was also a bit unimpressed with the snarkyness of the flyer, it should just be supported. Moshen Soussi from Afnic is thankful for the best effort support that he received, but ultimately a lot of people just have to get work done. He suggests using simpler feedback forms to see where the problems are. Don't rely too much on experts, just have normal people report their experiences.