Policy Proposal : "Run Out Fairly"
Dear colleagues, attached you will find a policy proposal we call "Run Out Fairly". It is based on a discussion instigated by your's truly during the open policy hour at the latest RIPE meeting. This is a proposal to gradually reduce the allocation and assignment periods in step with the expected life time of the IPv4 unallocated pool in order to address the perception of unfairness once the pool has run out. The proposal is not intended to stretch the lifetime of the unallocated pool. The proposal is independent of other proposals to reserve address space for transition purposes and/or new entrants. It can be implemented independently of these. We thank Filiz Yilmaz for her help, invite discussion of this proposal and look forward to high quality comments. For the authors Daniel Karrenberg PS: Can we refer to this proposal by name and not by number? ;-)
HI Daniel, On Mon, Apr 6, 2009 at 3:10 PM, Daniel Karrenberg <daniel.karrenberg@ripe.net> wrote:
Dear colleagues,
attached you will find a policy proposal we call "Run Out Fairly".
Seems fair to me, count me as +1 in support. -- Cheers, McTim
Hi all, This seems indeed fair and I therefor support this proposal. I have one question though, are there similair proposals out there in other regions ? Groet, MarcoH
Dear Marco, Yes, there are some proposals in other regions that seem to be similar to this one. Note that the similarity is often in the *intentions* rather than the detailed criteria they are proposing as of a solution to the foreseen problem. Below, please find a brief overview of these proposals: -o- In ARIN, there is a proposal called "Depleted IPv4 reserves". It is proposing to set a limit to be applied to all IPv4 address requests when ARIN's reserve of unallocated IPv4 address space drops below an equivalent of /9. By then, the address space that an organisation can receive will be limited to a single /20 within a six month period. This is currently under discussion and you can find further details at: https://www.arin.net/policy/proposals/2009_2.html -o- In APNIC, they have discussed two proposals in their last meeting back in February. One of them was to change the timeframe APNIC uses to make IPv4 allocations to meet LIRs' needs from twelve months to six months. The other was to set a maximum APNIC allocation size, by gradually decrease the maximum allocation size based on the remaining number of /8s in the IANA free pool. Both of these proposals were abandoned after APNIC 27, following a lack of community support. You can still find their details at: http://www.apnic.net/policy/proposals/prop-063-v002.html http://www.apnic.net/policy/proposals/prop-070-v001.html -o- LACNIC has reached consensus back in 2008 on a proposal called "Special IPv4 Allocations/Assignments Reserved for New Members". Accordingly when IANA's pool of IPv4 addresses is exhausted and LACNIC has a /12 left in their pool, only new members will receive address space from LACNIC. This space will be a /24 as the minimum and a /22 as the maximum. They have further detailed criteria and you can see them at: http://lacnic.net/documentos/politicas/LAC-2008-04-propuesta-en.pdf I hope this helps. Kind regards, Filiz Yilmaz Policy Development Officer RIPE NCC On Apr 7, 2009, at 1:14 PM, Marco Hogewoning wrote:
Hi all,
This seems indeed fair and I therefor support this proposal. I have one question though, are there similair proposals out there in other regions ?
Groet,
MarcoH
On 07.04 14:02, Filiz Yilmaz wrote:
...
In ARIN, there is a proposal called "Depleted IPv4 reserves". It is proposing to set a limit to be applied to all IPv4 address requests when ARIN's reserve of unallocated IPv4 address space drops below an equivalent of /9. By then, the address space that an organisation can receive will be limited to a single /20 within a six month period.
This is currently under discussion and you can find further details at:
While similar this appears to be a transition/new entrant reservation. It is orthogonal to "Run Out Fairly".
-o-
In APNIC, they have discussed two proposals in their last meeting back in February. One of them was to change the timeframe APNIC uses to make IPv4 allocations to meet LIRs' needs from twelve months to six months. The other was to set a maximum APNIC allocation size, by gradually decrease the maximum allocation size based on the remaining number of /8s in the IANA free pool.
Both of these proposals were abandoned after APNIC 27, following a lack of community support. You can still find their details at:
http://www.apnic.net/policy/proposals/prop-063-v002.html http://www.apnic.net/policy/proposals/prop-070-v001.html
These were the discussions I referred to. Note that one of the auhors of "Run Out Fairly" plays some role in the APNIC policy sig. ...
-o-
LACNIC has reached consensus back in 2008 on a proposal called "Special IPv4 Allocations/Assignments Reserved for New Members". Accordingly when IANA's pool of IPv4 addresses is exhausted and LACNIC has a /12 left in their pool, only new members will receive address space from LACNIC. This space will be a /24 as the minimum and a /22 as the maximum. They have further detailed criteria and you can see them at:
http://lacnic.net/documentos/politicas/LAC-2008-04-propuesta-en.pdf
Again more of a transition/new entrant reservation.
On 07.04 13:14, Marco Hogewoning wrote:
Hi all,
This seems indeed fair and I therefor support this proposal. I have one question though, are there similair proposals out there in other regions ?
Not yet, although I have seen some discussion in the same general direction. Of course it would not be ideal if only one region eventually did this. But this cannot be a reason to do nothing. If it turns out that at one point in the future we have too widely varying policies in effect across the regions, we have to review the situation and then again make the trade-off between different kinds of fairness. Groetjes Daniel
On 7 apr 2009, at 14:10, Daniel Karrenberg wrote:
On 07.04 13:14, Marco Hogewoning wrote:
Hi all,
This seems indeed fair and I therefor support this proposal. I have one question though, are there similair proposals out there in other regions ?
Not yet, although I have seen some discussion in the same general direction. Of course it would not be ideal if only one region eventually did this. But this cannot be a reason to do nothing. If it turns out that at one point in the future we have too widely varying policies in effect across the regions, we have to review the situation and then again make the trade-off between different kinds of fairness.
Agreed, we can't sit on our hands and wait. At the same time if the other 4 regions take a different approach this might encourage people to go shopping between RIR's. Groet, MarcoH
On 07.04 14:16, Marco Hogewoning wrote:
On 7 apr 2009, at 14:10, Daniel Karrenberg wrote:
On 07.04 13:14, Marco Hogewoning wrote:
Hi all,
This seems indeed fair and I therefor support this proposal. I have one question though, are there similair proposals out there in other regions ?
Not yet, although I have seen some discussion in the same general direction. Of course it would not be ideal if only one region eventually did this. But this cannot be a reason to do nothing. If it turns out that at one point in the future we have too widely varying policies in effect across the regions, we have to review the situation and then again make the trade-off between different kinds of fairness.
Agreed, we can't sit on our hands and wait. At the same time if the other 4 regions take a different approach this might encourage people to go shopping between RIR's.
Yep. And we have to deal with that then. Daniel
On Mon, Apr 6, 2009 at 2:10 PM, Daniel Karrenberg < daniel.karrenberg@ripe.net> wrote:
Dear colleagues,
attached you will find a policy proposal we call "Run Out Fairly". It is based on a discussion instigated by your's truly during the open policy hour at the latest RIPE meeting.
This is a proposal to gradually reduce the allocation and assignment periods in step with the expected life time of the IPv4 unallocated pool in order to address the perception of unfairness once the pool has run out.
The proposal is not intended to stretch the lifetime of the unallocated pool.
The proposal is independent of other proposals to reserve address space for transition purposes and/or new entrants. It can be implemented independently of these.
We thank Filiz Yilmaz for her help, invite discussion of this proposal and look forward to high quality comments.
For the authors
Daniel Karrenberg
PS: Can we refer to this proposal by name and not by number? ;-)
I have read the "Run Out Fairly" proposal and I support it. Ondrej -- Ondrej Sury technicky reditel/Chief Technical Officer ----------------------------------------- CZ.NIC, z.s.p.o. -- .cz domain registry Americka 23,120 00 Praha 2,Czech Republic mailto:ondrej.sury@nic.cz http://nic.cz/ sip:ondrej.sury@nic.cz <sip%3Aondrej.sury@nic.cz> tel:+420.222745110 mob:+420.739013699 fax:+420.222745112 -----------------------------------------
Hi, Can the authors please explain how this proposal would work with the current policy, which sets the minimum allocation at /21? For instance, on 2 January 2011, how much address space should be allocated to a network which needs a /21 for the next year? If the minimum allocation needs to change along with the allocation and assignment periods then that should probably be documented in the proposal. Thanks, Leo
On 07.04 11:11, Leo Vegoda wrote:
Hi,
Can the authors please explain how this proposal would work with the current policy, which sets the minimum allocation at /21?
For instance, on 2 January 2011, how much address space should be allocated to a network which needs a /21 for the next year? If the minimum allocation needs to change along with the allocation and assignment periods then that should probably be documented in the proposal.
Thanks,
Leo
I do not think a change is necessary here to address the objective of the proposal.
Dear Daniel, dear all, First of all I support this proposal, and thank you for taking the time to create it. I think the idea has great merit, but I¹m also reminded of an idea I sent out to the address policy mailing list and the feedback I got based on that. For that thread, see: http://www.ripe.net/ripe/maillists/archives/address-policy-wg/2008/msg00501. html . Just to refresh your memory, I proposed a policy that would only allocate a single block of space, regardless of the size of the request and available remaining inventory. One of the main shortcomings of my idea was that assignments from a new allocation don¹t happen in a gradual¹ way, which is one of the main assumptions behind any scheme based on time-windows. Larger organizations will just come back quicker not necessarily after the set window. I¹m afraid this proposal has the same weakness¹. Kind regards, Remco On 06-04-09 14:10, "Daniel Karrenberg" <daniel.karrenberg@ripe.net> wrote:
Dear colleagues,
attached you will find a policy proposal we call "Run Out Fairly". It is based on a discussion instigated by your's truly during the open policy hour at the latest RIPE meeting.
This is a proposal to gradually reduce the allocation and assignment periods in step with the expected life time of the IPv4 unallocated pool in order to address the perception of unfairness once the pool has run out.
The proposal is not intended to stretch the lifetime of the unallocated pool.
The proposal is independent of other proposals to reserve address space for transition purposes and/or new entrants. It can be implemented independently of these.
We thank Filiz Yilmaz for her help, invite discussion of this proposal and look forward to high quality comments.
For the authors
Daniel Karrenberg
PS: Can we refer to this proposal by name and not by number? ;-)
This email is from Equinix Europe Limited or one of its associated/subsidiary companies. This email, and any files transmitted with it, contains information which is confidential, may be legally privileged and is solely for the use of the intended recipient. If you have received this email in error, please notify the sender and delete this email immediately. Equinix Europe Limited. Registered Office: Quadrant House, Floor 6, 17 Thomas More Street, Thomas More Square, London E1W 1YW. Registered in England and Wales No. 6293383.
On 07.04 20:58, Remco van Mook wrote:
Dear Daniel, dear all,
First of all I support this proposal, and thank you for taking the time to create it. I think the idea has great merit, but I?m also reminded of an idea I sent out to the address policy mailing list and the feedback I got based on that. For that thread, see: http://www.ripe.net/ripe/maillists/archives/address-policy-wg/2008/msg00501. html . Just to refresh your memory, I proposed a policy that would only allocate a single block of space, regardless of the size of the request and available remaining inventory. One of the main shortcomings of my idea was that assignments from a new allocation don?t happen in a ?gradual? way, which is one of the main assumptions behind any scheme based on time-windows. Larger organizations will just come back quicker ? not necessarily after the set window. I?m afraid this proposal has the same ?weakness?.
Kind regards,
Remco
That can be so, but still the requests will be chopped up so that others can get in the queue rather than being pre-empted by a huge request. Daniel
Hi Daniel, That was the purpose of my idea as well this idea obviously is closer to currently set policy so the repercussions are likely to be better understood. It still leaves the option of cleaning out the cupboard¹ in one request, though. It just makes it harder to justify. (Which I¹m fine with, by the way, any extra limitation is better than none at all) Best, Remco On 08-04-09 16:40, "Daniel Karrenberg" <daniel.karrenberg@ripe.net> wrote:
On 07.04 20:58, Remco van Mook wrote:
Dear Daniel, dear all,
First of all I support this proposal, and thank you for taking the time to create it. I think the idea has great merit, but I?m also reminded of an idea I sent out to the address policy mailing list and the feedback I got based on that. For that thread, see:
http://www.ripe.net/ripe/maillists/archives/address-policy-wg/2008/msg00501.
html . Just to refresh your memory, I proposed a policy that would only allocate a single block of space, regardless of the size of the request and available remaining inventory. One of the main shortcomings of my idea was that assignments from a new allocation don?t happen in a ?gradual? way, which is one of the main assumptions behind any scheme based on time-windows. Larger organizations will just come back quicker ? not necessarily after the set window. I?m afraid this proposal has the same ?weakness?.
Kind regards,
Remco
That can be so, but still the requests will be chopped up so that others can get in the queue rather than being pre-empted by a huge request.
Daniel
This email is from Equinix Europe Limited or one of its associated/subsidiary companies. This email, and any files transmitted with it, contains information which is confidential, may be legally privileged and is solely for the use of the intended recipient. If you have received this email in error, please notify the sender and delete this email immediately. Equinix Europe Limited. Registered Office: Quadrant House, Floor 6, 17 Thomas More Street, Thomas More Square, London E1W 1YW. Registered in England and Wales No. 6293383.
Greetings! As I quickly read it I think it brings a fairness to the End. And it is good thing. But can you/somebody estimate what this would cost us in the matter of deaggregation and what it can cost NCC in the increase of request rate? I mean is it likely or at least possible, that when you shorten those periods from 1y to 1/2y and put in place the 50% usage in the half of the period criterion, then the increase of request rate (and therefore new routes in DFZ) will be not two-times but say ten-times more? In fact my concern is the same as in a section "Arguments Opposing the Proposal". My question is how much would we "pay" for the fairness? Best regards, Tomas Daniel Karrenberg wrote:
Dear colleagues,
attached you will find a policy proposal we call "Run Out Fairly". <cut> PS: Can we refer to this proposal by name and not by number? ;-)
-- Tomáš Hlaváček <tomas.hlavacek@elfove.cz>
* Daniel Karrenberg:
As of 1 July 2010, the RIPE NCC will start allocating enough address space to LIRs to meet their needs for a period of up to nine months.
As of 1 January 2011, the RIPE NCC will start allocating enough address space to LIRs to meet their needs for a period of up to six months.
As of 1 July 2011, the RIPE NCC will start allocating enough address space to LIRs to meet their needs for a period of up to three months.
This trades address space for routing table slots. Isn't this a bad thing? Do you suggesst to refuse requests for indepedently routable IP space if immediate need for the next three months does not justify a sufficiently large chunk of address space? -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99
Dear colleagues,
attached you will find a policy proposal we call "Run Out Fairly". It is based on a discussion instigated by your's truly during the open policy hour at the latest RIPE meeting.
I will support this policy. Jurek
On 06/04/2009 13:10, Daniel Karrenberg wrote:
attached you will find a policy proposal we call "Run Out Fairly". [...] PS: Can we refer to this proposal by name and not by number? ;-)
You'll have to forgive me for being suspicious about titles like this. It reminds me a little of the Democratic Peoples' Republic of Korea insisting on the word "democratic" being in their country's title, just to make sure that people are aware that it's a democracy - if they didn't already know. I'm curious to understand how smaller LIRs will cope with these requirements, particularly in light of Daniel's statement:
I do not think a change [of minimum allocation size] is necessary here to address the objective of the proposal.
If you're a small ISP, with a low customer signon rate, how are you going to justify requesting a /21 for 3 months usage when a /21 might normally last you for a period of, say 6 months or a year or something? I'm aware that the timescales are such that this period of enforced "fairness" won't be long. However it strikes me that during this period, very small ISPs will end up facing a tough fight to justify a /21 when it's clear from their previous allocation records that their address burn rate is very low and that /21 is significantly more than they would require within the assignment period. Nick
participants (11)
-
Daniel Karrenberg -
Filiz Yilmaz -
Florian Weimer -
Jerzy Pawlus -
Leo Vegoda -
Marco Hogewoning -
McTim -
Nick Hilliard -
Ondřej Surý -
Remco van Mook -
Tomas Hlavacek