Re: [sig-policy] A second policy proposal for the Policy SIG
Thanks for your comparison.
See below in-line some comments.
Regards,
Jordi
> De: TOYAMA Katsuyasu <toyama.katsuyasu at lab.ntt dot co dot jp>
> Responder a: <sig-policy-bounces at lists dot apnic dot net>
> Fecha: Fri, 02 Jun 2006 21:30:45 +0900
> Para: <sig-policy at apnic dot net>
> Asunto: Re: [sig-policy] A second policy proposal for the Policy SIG
>
> Dear SIG members
>
> I am Katsuyasu Toyama, who sent another proposal to policy SIG.
> This proposal is different from Jordi's proposal at the following
> points:
>
> (1) the target of portable assignment is limited only to end sites
> which are multihomed or plan to be multihomed, while Jordi does
> not restrict the target.
>
> End sites which want portability only for independence of
> service providers are NOT the target.
> The end sites which require redundancy with multiple links
> controled by BGP are the target because they have no other
> way to achieve it than portable assignment.
There are sites with have a single link but they have other reasons for
portable addressing. For example, an entity don't need to be multihomed
because SLAs warrantee enough their needs, but they may need to avoid
renumbering if exchanging to a new ISP.
We may enter here in the debate about what right a "so-called" critical
infrastructure is actually more critical than the independence for those
entities. One example can be a university, which has multiple campuses and
faculties and they are happy with a single link, and even they need a /32
because the number of sites, as they want to keep delegating a /48 to each
of one.
>
> I hope this can reduce unnecessary assignment, and avoid
> rapid expansion of the IPv6 global routing table.
I think that doesn't make a difference, those entities will have more
trouble, but will get in the routing table one way or the other.
>
> (2) assignment size is same as the size of non-portable assignment, /48,
> while Jordi's proposal is /32.
>
> The same size is better from the points of address consumption
> and migration from non-portable assignment to portable.
I think is in the other way around, is "more" portable a /32 in case this
entity in the future may opt to become an LIR.
>
> As I proposed, if assigned portable address space is separate
> from allocated portable, we can easily distinguish and filter
> the punching holes in allocated portable address space.
>
>
> Without the portable assignment for multihoming, I am afraid service
> providers may start to use punching hole for multhihoming on demand
> of customers, and it will cause messy situation just like IPv4.
>
This is going to happen anyway. The routing problem is a different one, of
much bigger impact.
>
> We are now preparing the comparison of Jordi's proposal, the proposal
> discussed in ARIN, and our proposal.
>
I'm working also in another proposal, also related to this, will try to
finish it during the weekend, so you may want to wait to compare that one
also ;-)
>
> Best regards,
> Katsuyasu Toyama
>
>
>
>
>
> Kenny Huang wrote:
>>
>>
>> Dear SIG members
>>
>> The proposal "IPv6 portable assignment for multihoming" has been sent to the
>> Policy SIG for review. It will be presented at the Policy SIG at APNIC
>> 22 in Kaohsiung, Taiwan, 4-8 September 2006. You are invited to review and
>> comment on the proposal on the mailing list before the meeting.
>>
>> The proposal's history can be found at:
>>
>> http://www.apnic.net/docs/policy/proposals/prop-035-v001.html
>>
>> Please feel free to submit your own policy proposal for discussion at APNIC
>> 22.
>>
>> Regards
>>
>> Kenny Huang
>> Policy SIG
>> huangk at alum dot sinica dot edu
>>
>>
>> ________________________________________________________________________
>>
>> prop-035-v001: IPv6 portable assignment for multihoming
>> ________________________________________________________________________
>>
>>
>> Authors: Katsuyasu Toyama
>> Takashi Arano
>> Tomohiro Fujisaki
>> Toshinori Ishii
>> Kosuke Ito
>> Dai Nishino,
>> Noriaktsu Ohishi
>> Izumi Okutani
>>
>> Version: 1
>>
>> Date: 2 June 2006
>>
>> SIG: Policy
>>
>>
>>
>> Introduction
>> ------------
>>
>> This policy allows 'end-sites' to be assigned IPv6 portable addresses only
>> if the end-sites are multihomed, or plan to be multihomed.
>>
>>
>>
>> Summary
>> -------
>> The current policy does not allow IPv6 portable assignment to any
>> end-sites. This obstructs end-site organizations which need redundancy
>> in internet connectivity for stable network operation.
>>
>> Shim6, another multihoming technology discussed in IETF, is not a
>> perfect replacement of the current multihoming technology using BGP due
>> to traffic engineering. In addition, it will take time to standardize
>> and implement Shim6.
>>
>>
>>
>> Situation
>> --------
>>
>> ARIN has been discussing the IPv6 Provider-independent address. The
>> draft was proposed in 2005 and moved to the last call after the meeting
>> consensus in April 2006.
>>
>> RIPE started PI discussion at RIPE in this May.
>>
>> AFRINIC and LACNIC discussed similar proposals recently in their Open
>> Policy meetings. In those regions, the issue has been returned to their
>> public mailing lists for further discussion.
>>
>> Note: APNIC uses the term "portable" rather than "provider-independent"
>> (PI).
>>
>>
>>
>> Details
>> -------
>>
>> (1) Assignment target:
>>
>> End-sites which are multihomed or plan to be multihomed, regardless
>> of their size.
>>
>>
>> (2) Assignment criteria:
>>
>> (2-a) The end site which is assigned IPv6 portable address space
>> must be multihomed using the assigned portable address space
>> in three (3) months.
>>
>> (2-b) If the portable address space is not used for multihoming
>> after three (3) months, the address space must be reclaimed.
>>
>> (2-c) The end site which is assigned IPv6 portable address space
>> must pay the fee for the space.
>>
>>
>> (3) Portable address space:
>>
>> (3-a) The portable assignment should be made from a specified block
>> separate from address space used for portable allocations
>>
>> (3-b) The portable assignment size to an end-site should be the same
>> size as in non-portable assignments, currently /48,
>> or a shorter prefix if the end-site can justify it.
>>
>>
>>
>> Pros/Cons
>> ---------
>>
>> Advantages:
>>
>> (1) Provides the solution for end-sites which require redundancy in
>> IPv6 and currently not able to do so due to the lack of
>> technical solutions.
>>
>> (2) Assignment of the portable address space is limited only to
>> 'multihoming purposes'; only end-sites which are or planned to
>> be multihomed can be assigned a portable address. This reduces
>> the consumption of portable address space as well as the growth
>> of the global routing table.
>>
>> (3) Portable assigned address space is separate from portable
>> allocated address space, therefore:
>>
>> (3-1) It helps preventing 'punching holes' in the portable
>> allocated address space because prefixes which are longer
>> than /32 can be filtered in portable allocated space.
>>
>> (3-2) it is relatively easy to abandon the portable assigned
>> address space in case some better techinical solutions
>> are developed in the future.
>>
>>
>> Disadvantages:
>>
>> It may lead to growth in the global routing table, but we think the
>> growth is almost the same in case that providers and end-sites start
>> using 'punching holes' for multihoming.
>>
>>
>>
>> Effect on APNIC
>> ---------------
>> No direct effect on the existing APNIC members, nor changes to the
>> current IPv6 allocation criteria.
>>
>>
>>
>> Effect on NIRs
>> --------------
>> NIR can adopt this policy at its discretion.
>>
>>
>> * sig-policy: APNIC SIG on resource management policy
>> *
>> _______________________________________________
>> sig-policy mailing list
>> sig-policy at lists dot apnic dot net
>> http://mailman.apnic.net/mailman/listinfo/sig-policy
>>
>>
>
> --
> TOYAMA, Katsuyasu <toyama.katsuyasu at lab.ntt dot co dot jp>
> IP Technology Expert Team, NTT PF-Labs.
> Nippon Telegraph and Telephone Corp.
> TEL: +81-(0)422-59-7906 FAX: +81-(0)422-59-5652
>
>
> * sig-policy: APNIC SIG on resource management policy
> *
> _______________________________________________
> sig-policy mailing list
> sig-policy at lists dot apnic dot net
> http://mailman.apnic.net/mailman/listinfo/sig-policy
**********************************************
The IPv6 Portal: http://www.ipv6tf.org
Barcelona 2005 Global IPv6 Summit
Slides available at:
http://www.ipv6-es.com
This electronic message contains information which may be privileged or confidential. The information is intended to be for the use of the individual(s) named above. If you are not the intended recipient be aware that any disclosure, copying, distribution or use of the contents of this information, including attached files, is prohibited.