Re: [sig-policy] IPv4 countdown policy proposal
On Wed, 14 Feb 2007 09:04:35 +0900 Takashi Arano wrote:
> If someone can propose a graduated approach, then we can compare it
> with the proposed policy.
Firstly let me apologise if I have missed something in this debate on
the proposed policy and well done to the authors for addressing this
critical and intractable issue.
I see the proposals for Global synchronization as very attractive and
hopefully progress can be made here.
Like David Conrad I believe a graduated approach to rationing would be a
more appropriate response than maintaining current policies. A
gradual(?) tightening of the allocation criteria over the next few
years, as certain promulgated thresholds are passed, would have three
principal effects:
a) It would extend the life of the remaining pool (making the proposal
to retain blocks of v4 address space against some unanticipated need
unnecessary) and
b) It would provide more certainty on allocation principals in the run
up to ultimate exhaustion and,
c) Provide a motivation to ISP's to do something with v6 (which for all
it's problems seems to be the only long term solution on offer).
Remaining with existing policies until exhaustion (real or artificial)
would, I suspect, not only lead to chaos for those whose countries/
businesses/ customers require new allocations but also generate an
active trading market (either black or otherwise) in allocations.
At the moment there is no incentive for ISP's to adopt potential
solutions to v4 exhaustion such as dual v4/v6 stack as the deployment
costs are relatively high and the benefits for an individual ISP are, at
best, negligible.
I'd suggest that the countdown policy should include measures to first
promote and later require some of the steps for an orderly transition to
the deployment of v6.
In other words a first step could be (say) a requirement for some
deployment of v6 in order to qualify for further allocations of v4
space. This measure alone would raise the bar on obtaining allocations
and require some of the real problems of migration to v6 to start being
addressed.
Clearly it would be difficult to enforce large v6 deployment but as a
first step it should be relatively easy to devise a simple test that a
member has some dual stack connectivity?
Other measures such as trading and recycling of v4 space are desperately
required but should not delay the development of a countdown roadmap.
--
Robert Gray
bob at brockhurst dot co dot nz