Draft Minutes (v1) of IPv6 WG session RIPE51
Please see enclosed the first version of the minutes of our session at RIPE51. I plan to declare the minutes final if I don't receive any comments for significant changes by November 4th 2005. I would like to thank Timothy Lowe from the RIPE NCC for taking the minutes. David Kessens --- Minutes IPv6 Working Group session Amsterdam, 12 Oct, 2005 Chair: David Kessens Scribe: Timothy Lowe # of attendees: 95 ACTION 51.1 - Ask the RIPE NCC to revise the acronym list in the registration package to remove obsolete terms as TLA & NLA (David Kessens) Status: Completed ACTION 51.2 - Create a link from the IPv6 WG page to the APNIC events page: http://apnic.net/info/calendar/index.html (RIPE NCC) Status: To be done ACTION 51.3 - Provide traffic reports on ipv6 usage during the RIPE meetings. If available, data from earlier meetings will be made available as well. (RIPE NCC) A. Administrative stuff Appointment of scribe: Timothy Lowe Agenda bashing: No additions to agenda Action point 51.1 was decided. B. Quick update from the RIPE NCC regarding ipv6 services (RIPE NCC) No news from the RIPE NCC C. IPv6 events calendar on the ipv6 working group page (proposed by Timothy Lowe) It was decided that similar event pages already exist and that we can simply provide a link to the APNIC event pages. (See action point D. Discussion on: Global IPv6 routing table status (Gert Doering, Spacenet) Discussion on filtering: Gert Doering: The RIPE Whois database is the only one allowing route6 object creation. Iljitsch Van Beijnum: What is the gain of filtering less specific ? Geoff Huston: What problem are you trying to solve ? Gert Doering: Trying to prevent accident, less maintenance if the documentation is in the database then it will be more stable. Conclusion: as we cannot force people to filter, we will see more specifics anyways (David Kessens) E. Report(s) about *actual* v6 traffic volume as compared to v4? *what's real* out there, not what's on powerpoint? - http://www.sixxs.net/misc/traffic/ (Jeroen Massar) Pointer to Jeroen's website was provided, but content was not discussed. Pim van Pelt offered to do some work to provide more data on his Sixxs ipv6 service but it was decided that the amount of work was probably not worth the effort. Discussed whether any ipv6 usage data from the RIPE meetings itself was available. This discussion resulted in a request and an action point for the RIPE NCC to provide us such data in the future (Action point 51.3) F. The end of the 6bone address space is in sight (Jeroen Massar) Discussion: - what are people going to do with 6bone address space as - per 6/6/6 ? - filter it? - keep it going? - warn people? - what to do with the 6bone registry ? Iljisch v B.: DNS is not auto-configurable in IPv6 like it is in IPv4 so I would like to use 6bone to get an ipv6 assignment for an experiment for well-known ipv6 DNS resolver addresses. Geoff H. : IANA has released a statment that 6bone dies on 06/06/2006. Gerd D.: exactly, what will people do when this happens? No uniform policy on what to do with 6bone prefixes. Some people will start filtering while others won't bother. In general, people advised David Kessens who maintains the 6bone registry to stop accepting registry changes by the close date but to keep the data online for some time to come with proper warning message when information is accessed. G. Developments/initiatives regarding IPv6 in the RIPE region and beyond - 6DISS (Bernard Tuy) Presentation will be available on the RIPE website as soon as Bernard Tuy provides the slides No questions. H. Input for the RIPE NCC Activity Plan (no input from the audience) Z. AOB Peter Koch drew attention to a presentation in the DNS working group regarding the deprecation of IP6.int. ----
[Microsoft/OpenTransit folks, look for '3ffe:8310::/28' for the Teredo prefix question] On Fri, 2005-10-21 at 06:01 -0700, David Kessens wrote:
Please see enclosed the first version of the minutes of our session at RIPE51.
I plan to declare the minutes final if I don't receive any comments for significant changes by November 4th 2005. <SNIP>
Just a couple of notes, thus no comments on the minutes themselves, as I was not present, thus I might repeat some obvious things that where already said during the meeting. {There are at least two typo's look for the braces}
Conclusion: as we cannot force people to filter, we will see more specifics anyways (David Kessens)
What I have seen is that a number of times, when more specifics are announced, eg /48's that the 'better quality' networks filter them out, the prefix will then be carried only by 'less qualitity' networks or more politically correct, the route goes a couple of times around the world. Notorious are mostly Korean 'transit providers' who seem to accept anything. <SNIP> {NOTE: the typo in IljiTsch's name on the next line}
Iljisch v B.: DNS is not auto-configurable in IPv6 like it is in IPv4 so I would like to use 6bone to get an ipv6 assignment for an experiment for well-known ipv6 DNS resolver addresses.
Iljitsch, for this there is of course: http://www.ripe.net/ripe/docs/ipv6policy.html#experiment-assignments No need for 6bone space :) Next to that, dhcpv6 of course works eg using dibbler, which provides the same mechanism. I am also in favor of the idea that Iljitsch has and to have a single IP address (which has to be part of a /32 or so to be not filtered...) which is the 'closest' dns service. I tried this in sort of a way at the last IETF though, but the word 'anycast' freaks out quite some people, see section 4 of the following for the relevant part: http://unfix.org/~jeroen/archive/drafts/draft-massar-dnsop-service-00.txt Though I understood that that doc hit on about 5 holy wars or so :)
Geoff H. : IANA has released a statment that 6bone dies on 06/06/2006.
{NOTE: GerT with a T}
Gerd D.: exactly, what will people do when this happens?
No uniform policy on what to do with 6bone prefixes. Some people will start filtering while others won't bother.
From RFC3701: 8<--------------- Thus after the 6bone phaseout date June 6, 2006, it is the intent that no 6bone 3FFE prefixes, of any size/length, be used on the Internet in any form. Network operators may filter 3FFE prefixes on their borders to ensure these prefixes are not misused. --------------->8 Nothing states though how to handle the case where people are announcing these prefixes, my intention is to start contacting people who are currently still using their prefix and reminding them that they need to drop it. This is the same situation when some ISP simply starts announcing eg 2005:1234:/32, what happens? People accept it or? Provided one has enough resources (read: cash) I actually guess most ISP's can be convinced/bribed that they will be accepted and IANA/RIR's are out of the loop. Then again remembering the 2003::/16 prefix from Microsoft, that got filtered quite hard by many, though I guess also because of some 'anti-m$' thoughts there ;) Which leads to another nice question, Teredo's are per default using a 6bone prefix (3ffe:8310::/28).... what is happening with that one because I assume that won't disappear from this planet 1,2,3!? At the moment that prefix is being announced by Microsoft (AS8070) The 3ffe:831f::/32 is also announced by OpenTransit (AS5511): inet6num: 3FFE:831F::/32 netname: WANADOO-EXP-IPV6-6BONE descr: France Telecom Wanadoo IPV6 experimentations country: FR admin-c: ML50-6BONE tech-c: ML50-6BONE mnt-by: FT-BRX changed: gestionip.ft@francetelecom.com 20050721 source: 6BONE That date is pretty new btw, clearly from GRH most ISP's also accept it. When folks got 6bone pTLA's they also agreed to read the 6bone mailinglist, but it has been seen already quite some times that many don't, directly contacting is thus going to be required. http://www.sixxs.net/tools/grh/dfp/6bone/ 8<---------------------- The database currently holds 144 IPv6 DFP's. Of which 31 (21.53%) are returned to the pool, 29 (20.14%) IPv6 DFP's didn't have a routing entry. Thus 84 (58.33%) networks are currently announced. ---------------------->8 about 60% left... Another related note on this subject; what is going to happen to www.6bone.net ? Is it going to be updated to something nice referring to information about IPv6 or? There are also quite a number of broken links on that page and most of the information is outdated or has been revised.
In general, people advised David Kessens who maintains the 6bone registry to stop accepting registry changes by the close date but to keep the data online for some time to come with proper warning message when information is accessed.
The registry is a good thing to keep, it is like having an old phone book around and your phonebills from that era, one can lookup the 'historical' data in there. So I agree with the above. Greets, Jeroen PS: I did a presentation about SixXS / Deploying IPv6 at SwiNOG #11, for interrested people, see: http://www.swinog.ch/meetings/swinog11/ Misses the words and it is indeed quite big but pictures tell quite a bit.
participants (2)
-
David Kessens -
Jeroen Massar