You're viewing an archived page. It is no longer being updated.
RIPE 48
RIPE Meeting: |
48 |
Working Group: |
Address Policy |
Status: |
Final |
Revision Number: |
1 |
Address Policy Working Group Summary (RIPE 48)
------------------------------------------
Date: Wednesday 5 May 2004
Time: 09.00 - 12.30
Location: Grand Ballroom
------------------------------------------
Agenda
A: Administrative matters
B: RIPE NCC (stats) Update
C: ICANN ASO AC Update
D: ICANN ASO AC - A view from outside
E: Discussion resulting from ICANN ASO presentations
F: Looking at the 80% Policy for Additional IPv4 Allocations
G: Policy Development Process
H: Issues seen in IPv6 Allocation Requests
I: IPv6 Policy Discussion
J. Expanded Address Allocation for Private Internets
Y: Open Microphone
Z: AOB
------------------------------------------
A: Administrative matters
Chair: Hans-Petter Holen
Co-Chair: Gert Doering, Andrea Borgato
Scribe: Laura Cobley, Scott Donald
Minutes from RIPE 47 approved:
Update on open actions:
B: RIPE NCC (stats) Update - Filiz Yilmaz (RIPE NCC)
Leslie Nobile (ARIN) informed us that the HD Ratio idea was rejected in the ARIN region and clarified that the Address Policy for multihomed networks does not only apply ISPs, but also includes End Users
The terminology [slide 3] was taken directly from the IANA website and can be updated in co-ordination with IANA:
Private Use = 10/8 to be reserved for private use.
Public Data = 14/8 for public networks - x25s
Central Registry = pre-RIR address space (old "Class A" & ERX
space)
There was a call for the RIPE NCC to take a more active role here, as stated by the NRO. Axel Pawlik (RIPE NCC) commentedm that the RIPE NCC wants to work closely with IANA concerning these figures and will make recommendations based on the feedback from the community.
C: ICANN ASO AC Update - Hans Petter Holen
The ASO representatives participate in global policy development. If anyone has any issues they would like us to bring up, please let us know.
D: ICANN ASO AC - A view from outside - Hans Petter Holen
We were unable to find a volunteer for this - perhaps we can add it to the agenda for the next meeting.
E: Discussion resulting from ICANN ASO presentations
no comments
F: Looking at the 80% Policy for Additional IPv4 Allocations - Paul Wilson (APNIC)
(AKA HD Ratio for IPv4)
Paul Wilson (APNIC) stressed that the current 80% rule is not based on any scientific facts or principles any more than the HD ratio method. The idea has already been presented to the APNIC community to raise awareness and received a fairly positive reaction.
There were some concerns that administrative staff would not be able to understand the system and it would make things much more difficult administratively as well as effectively 'wasting' a lot more address space.
Leo Vegoda (RIPE NCC) stated that as the RIPE NCC is always seeking to minimise any burden on its members, it would be prepared to write software alleviating this administrative burden. There would be no need to worry about the calculations.
John Curran (ARIN) mentioned that in the ARIN region, a lot of people didn't understand the value of the system and didn't see the problem that was trying to be solved.
It was generally agreed that the issue of worldwide shopping for address space (global ISPs with international networks) is a concern that needs to be carefully considered before adopting different procedures in each region. Maybe the NRO should be asked to step in.
Action (WG Chair) Take discussion to the mailing list
H: Issues seen in IPv6 Allocation requests - Leo Vegoda (RIPE NCC)
There were some concerns that we were concentrating on clarifying the text instead of working together with the other Regions on this. It was reiterated that the RIPE NCC is working together with the other RIRs but wanted to highlight the issues that will need further discussion within the working group.
There was clarification on the paragraphs relating to privacy laws (the document had to apply on a global level) and registration (any organisation holding more than one /48, contiguous or not, needs to register this). This type of information needs to be clarified in the policy if only a handful of people know.
I: IPv6 Policy Discussion - Gert Doering (SpaceNet AG)
Louis Lee (Equinix, Inc.) commented that as a US operator, if asked to supply a plan, this is considered to be a marketing forecast. If this is not the case, the wording in the policy needs to clearly state what is required.
There were calls for clarification of the initial allocation criteria and possibly to review the 200 /48 requirement.
Some organisations (such as start-up companies) think that unless they have solid plans to have 200 users they don't qualify. They don't realise that they are only required to think that they will have 200 users in the foreseeable future.
There was also some concern over site-locals and whether they
can be used for private address space, however this is being
looked at by the IETF who are working on a form of unique site
locals that would be unique, but not routable. This work has
been neither passed nor rejected so far.
Leslie Nobile (ARIN) mentioned that in the ARIN region the requirements in section 5.1.1.d have been changed to "Having 200 /48 assignments in 5 years OR be a "known" ISP in the ARIN region"
There was some confusion over the minumum allocation size (/32), but it was clarified that as long as an LIR can show that they need an initial allocation, the policy provides for this.
It was also clarified that the decision to restricting the assignment sizes to /128, /64 and /48 (RFC3177) was taken to alleviate possible renumbering issues (i.e. /52 to /48) and due to the fact that there are more than 1000 /48s for every person on the planet.
Jordi Palet (Consulintel) expressed concern that there are ISPs that don't understand the assignment recommendations and refuse to assign more than a /64 to their customers.
It was confirmed that this is a misconception on the ISPs part and it was also clarified that you don't have to be an LIR to obtain an experimental allocation, however the request does need to be sent by an LIR.
There was some uncertainty over what size assignment LIRs should give to End-customers for End-site multihoming. This might be more relevant for the Routing WG but it seems to be working ok with most /40 and /48s being advertised. The whole issue is fairly open at the moment.
Another suggestion was that LIRs should be prevented from holding address space if they don't use it within a certain time period. It already seems [from the slides] that about 30% of the allocated address space is not advertised. On the other side, this might put people off when the use of IPv6 should be encouraged. The original concern with the policy had been to avoid adding to the size of the routing table.
Wilfried Woeber (Vienna University) confirmed that he wasn't too worried about abolishing the initial allocation requirements completely. Those people that want to announce them can do so, and those that don't, don't.
There seem to be mixed feelings about whether this is too radical.
Leo Vegoda (RIPE NCC) brought up some concerns from the EIX WG, whereby simply allocating a /32 to all LIRs would be preferable so that IXPs that are LIRs can provide for their services (not for their network mesh).
The only question remaining is how to coordinate with the other regions. Perhaps the RIPE NCC will do this?
Action (WG Chair) - Circulate input on the mailing list.
Action (RIPE NCC) - Propose re-write following mailing list discussion.
J: Expanded (IPv4) Address Allocation for Private Internets - Gert Doering (SpaceNet AG)
http://www.ietf.org/internet-drafts/draft-hain-1918bis-00.txt
It was suggested to use class E space for this, but there were other concerns that this address space would only be routed by BSD.
G: Policy Development Process - Hans Petter Holen (Tiscali)
no comments
Y: Open Microphone
no comments
Z: AOB
no 'other business', so the meeting was closed.
------------------------------------------