Dear IPv6 WG list members, you can find the draft minutes from the IPv6 WG session held during RIPE 89 in Prague on the website [1]. Thanks to the RIPE NCC for taking them! If you have any comments or questions, please let us know. [1] https://www.ripe.net/community/wg/active-wg/ipv6/minutes/ipv6-working-group-... --------------------CUT-------------------- IPv6 Working Group Minutes - RIPE 89 ==================================== Date: Thursday, 31 October 09:00 - 10:30 (UTC+1) Chairs: Christian Seitz, Nico Schottelius, Raymond Jetten Scribe: Emile Aben Status: Draft View the recordings (https://ripe89.ripe.net/programme/meeting-plan/ipv6-wg/) View the stenography transcript (https://ripe89.ripe.net/archives/steno/40/) View the chat logs (https://ripe89.ripe.net/archives/chat/40) 1. Welcome ---------- The presentation is available at: https://ripe89.ripe.net/archives/video/1491 Co-chair Christian Seitz opened the session. 2. Why Is This IPv6 Transition Taking So long? ---------------------------------------------- Geoff Huston, APNIC The presentation is available at: https://ripe89.ripe.net/archives/video/1493 Wai Sing Samson Chooi (Messrs. Chang Chooi & Co.) asked if the RIPE NCC had resources to support capacity building for new members. Raymond Jetten (co-chair) responded that the RIPE NCC had quite a few training available. Kurt Kayser (Kurt Kayser Konsultation) objected to comparing IPv6 to steam trains and asked Geoff to change the analogy. The room applauded Kurts’ suggestion. Geoff acknowledged the suggestion. Steve Wallace (Internet2) brought up the 2,600 ASNs that make up the global Research and Education Networks as a counter example. Geoff answered that research and commercial networks have diverged, and that scaling properties of address-based architectures was really difficult and expensive. Jen Linkova (Google) completely disagreed with Geoff that NAT is free as she was dealing with users that were unhappy about it. She agreed that the problem with the transition was human nature, as many only acted if it was really needed. Geoff replied that IPv4 could only be turned off when it was no longer in use and that, from an economic point of view, there was no momentum for the transition. 3. K8s IPv6 and Routing ----------------------- Nico Schottelius, ungleich The presentation is available at: https://ripe89.ripe.net/archives/video/1497 Jan Zorz (6connect) asked if you could do VPP DPDK in these containers. Nico replied that it was possible. Dmytro Shypovalo (Vegvisir Systems) asked if Nico explored SRv6 with Kubernetes. Nico said that he did not. Advent of IPv6 Jens Link The presentation is available at: https://ripe89.ripe.net/archives/video/1498 Ondřej Caletka (RIPE NCC) liked the idea of an IPv6 advent calendar and said the RIPE NCC has content to contribute. He offered to help. Jens then asked the room if it was a good idea or not. The feeling in the room was that it was a good idea. Jens asked who would be willing to volunteer and got only a few hands up. Jen Linkova (Google) offered the idea of giving people pcaps to take a look at what’s wrong with them. Jens thought that was a good idea. Raymond concluded that Jens’ idea looked like a good idea. 4. What We Have Achieved So Far ------------------------------- Benedikt Stockebrand (stepladder IT) The presentation is available at: https://ripe89.ripe.net/archives/video/1500 Jen Linkova (Google) said that there was no point in telling people what to do, so she disagreed with the advice and pointed out that we needed technical solutions to stop bad things happening. She asked if Benedikt had any figures on the increasing number of routes that made him so concerned about the routing table. Benedikt replied that he had anecdotal evidence. Jen said that she had seen large deaggregation events, but that those were mistakes and could be mitigated by setting max prefix limit. She suggested Benedikt to go into more technical details on how to prevent bad things from happening. Benedikts' advice was to remain alert and that it was difficult to give advice beyond monitoring and being prepared. Yannis Nikolopoulos (CellMobile) asked to what extent we have failed with transition technologies. Benedikt replied that they prolonged the pain because they are difficult to use and cost a lot of money. Gert Doring (Spacenet) replied that transition technologies were in fact extremely well made. Gert then gave the advice to filter customers for what they intend to announce. Benedikt responded that he suspected a serious impact due to new entrants that are not ISPs in the next couple of years. Regarding routing and disaggregation, Gert said that people should filter their clients. He added that this was something that all transit networks should be doing, and that most people in the room were already doing, so this was not the right forum. Benedikt said the key point was to have a rush of non-ISP players, and he suspected a serious impact in that regard. 5. Wrap Up ---------- This part of the session is available at: https://ripe89.ripe.net/archives/video/1501/ Ondřej Caletka (RIPE NCC) had an AOB. He asked for NetworkManager to support IPv6-only networks. He specifically asked paying customers of the product to make this request. Raymond then closed the session. --------------------CUT-------------------- Thanks and kind regards Chris -- Individual Network Berlin e.V. : support@in-berlin.de : vorstand@in-berlin.de Tel +49-30-45494343 ::: Fax +49-30-45494344 ::: Web https://www.in-berlin.de/ IN-Berlin e.V. : Christian Seitz (1. Vors.) : Lehrter Str. 53 :: 10557 Berlin Amtsgericht Charlottenburg 95 - VR 15669 Nz ::::::: USt.Ident-Nr. DE188894648