Allocations from the last /8

Summary of Proposal:

This proposal describes how the RIPE NCC should distribute address space from its last /8 worth of IPv4 address space.

a. Arguments supporting the proposal

The final /8 worth of address space will have a special policy applicable to it in the RIPE region. This avoids the risk of one or a few organisations consuming the entire block with a well crafted and fully justified resource application. The proposal attempts to ensure that no organisation lacks real routable IPv4 address space during the coming transition to IPv6. It is very important that new entrants also have the chance to get some IPv4 space before it totally runs out because a small block of IPv4 addresses will be necessary for a while before the Internet can run on an IPv6-only platform.

b. Arguments Opposing the Proposal

IPv4 PI addresses will not be assigned through this policy.

Some organisations may believe and can demonstrate that their IPv4 requirements are larger than the minimum allocation or assignment sizes in RIPE region. But this final /8 is not intended as a solution to the growth needs of a few organisations, but for assisting with the transition from IPv4 to IPv6.

Some organisations may set up multiple LIR registrations in an effort to get more address space than proposed. The RIPE NCC must be vigilant regarding these, but the authors accept that it is hard to ensure complete compliance.

Note: In order to provide additional information related to the proposal, details of an impact analysis carried out by the RIPE NCC are documented below. The projections presented in this analysis are based on existing data and should be viewed only as an indication of the possible impact that the policy might have if the proposal is accepted and implemented.

A. RIPE NCC's Understanding of the Proposed Policy

The proposed policy will be triggered once the RIPE NCC holds the equivalent of a /8 of unallocated and/or unassigned IPv4 Address Space. This applies solely to the address space that has been allocated to the RIPE NCC by the IANA. Address space blocks with legacy assignments or allocations are outside the scope of this proposal.

This address space may not be a single aggregated /8 and can contain prefixes as long as a /29 or more. Therefore it is possible that at some point the RIPE NCC will be unable to allocate a single /22 block to an LIR and will have to allocate multiple prefixes that total 1,024 IP addresses. Because of this situation, there will effectively no longer be a minimum allocation size within the RIPE NCC service region when this happens.

It is possible that the final request for an allocation that causes this policy to be triggered may only be partially fulfilled by the RIPE NCC before the threshold of the last /8 is reached.

Under the current policies, when this proposed policy is triggered, the RIPE NCC will no longer assign ASSIGNED PI and ASSIGNED ANYCAST space. If proposal 2010-01, “Temporary Internet Number Assignment Policies”, is accepted we will still assign temporary ASSIGNED PI space from a separate pool that we will reserve for this purpose.

It is possible that some IPv4 Address Space will be returned to the RIPE NCC even after this proposed policy has been triggered. Any such returned space will be added to the pool that is available for /22 allocations to LIRs under this proposed policy, even if the returned space causes the unallocated pool to rise to more than a /8 in total.

B. Impact of Policy on Registry and Addressing System

Address/Internet Number Resource Consumption:

If accepted, this proposed policy will slow down the consumption of the last IPv4 addresses because each LIR will receive only one allocation and no PI assignments will be made from the last /8.

Currently, the RIPE NCC has about 7000 members. The equivalent of a /8 will accommodate 16,384 allocations of a size /22. The /16 set aside as a reserve reduces the number by 64, leaving a potential of 16,320 allocations to be made under this proposed policy. Thus the proposal accommodates for explosive growth of the RIPE NCC membership or could enable the RIPE NCC to have IPv4 space for many years to come.

Fragmentation/Aggregation:

As mentioned above, there is possibility that the last /8 may not be a single aggregated /8 and may contain prefixes as long as a /29. However, for the sake of this analysis we have looked at the case of a block where the longest prefix will be a /22.

Ultimately, once the reserved /16 is used, the proposal splits the last /8 into 16,384 allocations of size /22. This is significantly higher than the fragmentation observed in any of the /8s which have been allocated to the RIPE NCC before. Most of these /8s have less than 1,000 allocations and assignments (source: ftp://ftp.ripe.net/pub/stats/ripencc/delegated-ripencc-latest).

Compared to the current global routing table statistics as reported in http://www.cidr-report.org/as2.0/, the number of prefixes would go up by 5% when all 16,384 allocations have been handed out. This growth could come in two stages. In the first stage we could see a rapid increase in a relatively short time when the threshold of the last /8 is reached and existing LIRs matching the criteria each apply for a /22. In the second stage, we could see the extra growth happen more gradually while new entrants join the RIPE NCC and apply for their /22 of IPv4.

C. Impact of Policy on RIPE NCC Operations/Services

The policy proposal requires that the RIPE NCC is vigilant regarding those organisations that set up multiple LIRs in order to circumvent the "one IPv4 allocation per LIR" restriction. Although the authors accept that it is hard to ensure complete compliance, a level of compliance is required. At the moment the RIPE NCC has no mechanisms to control the number of LIRs per organisation. To comply with the proposed policy, such control mechanisms must be developed.

D. Legal Impact of Policy

This policy proposal has many similarities with the withdrawn policy proposals 2008-06  and 2009-04
The RIPE NCC has studied the legal implications of these previous proposals. The conclusions of this study have been published on the Address Policy Working Group mailing list:

http://www.ripe.net/ripe/maillists/archives/address-policy-wg/2009/msg00797.html

and have been presented during the RIPE 59 Meeting:

http://www.ripe.net/ripe/meetings/ripe-59/presentations/pawlik-final-shalsh-8.pdf

On the basis of this study, the RIPE NCC does not anticipate that the implementation of this proposed policy would cause any significant legal implications.