Re: [address-policy-wg] 2011-05 New Policy Proposal (Safeguarding future IXPs with IPv4 space)
On 25/10/2011 10:51, Emilio Madaio wrote:
Dear Colleagues,
A proposed change to RIPE Document ripe-530, "IPv4 Address Allocation and Assignment Policies for the RIPE NCC Service Region", is now available for discussion.
I hate to be a killjoy here, but doesn't the average IXP use less than a /22? Part of the rational says: "Much of the work that goes into creating Internet Exchange Points today is done in regions of the world where there is no Internet Exchange Point yet. Opening an IXP helps to develop the Internet and Internet community in that region, so the work really is 'good of the Internet'. Starving these regions of an opportunity to build a simple open Internet Exchange Point would be severely damaging to Internet users in regions under-served by IXPs." But a brand new IXP will be able to get a /22 under the existing rules. Surely that's enough? I have a general dislike of "special cases" especially where they don't seem to be necessary. And this seems to be unnecessary. Nigel
On 27 Oct 2011, at 13:25, Nigel Titley wrote:
But a brand new IXP will be able to get a /22 under the existing rules. Surely that's enough?
I have a general dislike of "special cases" especially where they don't seem to be necessary. And this seems to be unnecessary.
+1 on both points. I wonder too about how this policy could be gamed. What happens if new (or existing) LIRs pretend to be IXPs to snatch extra v4 space they wouldn't otherwise get? Once the v4 ship really starts to sink, people are bound to grap hold of anything that looks like a lifeboat and make sure nobody else gets to climb aboard.
Hi Jim -- On 27 Oct 2011, at 14:26, Jim Reid <jim@rfc1035.com> wrote:
I wonder too about how this policy could be gamed. What happens if new (or existing) LIRs pretend to be IXPs to snatch extra v4 space they wouldn't otherwise get?
The IPRAs were very diligent enough to prevent this from happening when IXPs were allowed v6 PI, when all other organisations were not. There was a clean definition written which has stood firm in later analysis in ap sessions. I think this risk is well mitigated against. Andy
On 27 Oct 2011, at 17:54, Andy Davidson wrote:
I wonder too about how this policy could be gamed. What happens if new (or existing) LIRs pretend to be IXPs to snatch extra v4 space they wouldn't otherwise get?
The IPRAs were very diligent enough to prevent this from happening when IXPs were allowed v6 PI, when all other organisations were not.
There was a clean definition written which has stood firm in later analysis in ap sessions.
I think this risk is well mitigated against.
OK. Thanks Andy. I raised this as a theoretical concern. So provided the IPRAs can keep a lid on potential abuse of the policy, that's (sort of) fine. That'll be "good enough" IMO. I don't have a strong objection to the proposal even though I still have some reservations.
Hi there, On 27 Oct 2011, at 19:05, Jim Reid <jim@rfc1035.com> wrote:
OK. Thanks Andy. I raised this as a theoretical concern. So provided the IPRAs can keep a lid on potential abuse of the policy, that's (sort of) fine. That'll be "good enough" IMO. I don't have a strong objection to the proposal even though I still have some reservations.
Thanks for helping to give the policy proper scrutiny and for your support ! Best wishes, Andy
On 27 Oct 2011, at 13:25, Nigel Titley wrote:
Part of the rational says: "Much of the work that goes into creating Internet Exchange Points today is done in regions of the world where there is no Internet Exchange Point yet. Opening an IXP helps to develop the Internet and Internet community in that region, so the work really is 'good of the Internet'. Starving these regions of an opportunity to build a simple open Internet Exchange Point would be severely damaging to Internet users in regions under-served by IXPs."
But a brand new IXP will be able to get a /22 under the existing rules. Surely that's enough?
Hi, Nigel and Jim Thanks for the emails. Almost no IXPs have PA, and most community based ones don't have budget for their own LIR. This policy allows successful IXPs to be born in 2012 as they have been between 1994 and 2011. i.e. the equivalent of PI. It will provide policy parity with IPv6 too, where IXP specific policy exists in order to sidestep restrictions designed for other networks. Thanks Andy
Thanks for the emails.
Almost no IXPs have PA, and most community based ones don't have budget for their own LIR. This policy allows successful IXPs to be born in 2012 as they have been between 1994 and 2011. i.e. the equivalent of PI. It will provide policy parity with IPv6 too, where IXP specific policy exists in order to sidestep restrictions designed for other networks. Hmm, yes, I see what you mean. I think what might have confused me is
On 27/10/2011 15:56, Andy Davidson wrote: the 5.6 section heading "Use of last /8 for PA allocations". The IXP allocation section (not being PA, as you've pointed out) doesn't really belong here. And yes, I agree this is nit-picking, before anyone says it for me. And following the clarification I'm happy with the intent of the proposal (whilst still being unhappy with special cases in general). Nigel
Hi there, On 27 Oct 2011, at 16:42, Nigel Titley <nigel@titley.com> wrote:
On 27/10/2011 15:56, Andy Davidson wrote:
Thanks for the emails.
Almost no IXPs have PA, and most community based ones don't have budget for their own LIR. This policy allows successful IXPs to be born in 2012 as they have been between 1994 and 2011. i.e. the equivalent of PI. It will provide policy parity with IPv6 too, where IXP specific policy exists in order to sidestep restrictions designed for other networks. Hmm, yes, I see what you mean. I think what might have confused me is the 5.6 section heading "Use of last /8 for PA allocations". The IXP allocation section (not being PA, as you've pointed out) doesn't really belong here.
I see. You are right, well spotted. Maybe we should vary the section headed to "Uses of the final /8" ?
And yes, I agree this is nit-picking, before anyone says it for me.
And following the clarification I'm happy with the intent of the proposal (whilst still being unhappy with special cases in general).
Thank you for your support and good ideas. Andy
participants (3)
-
Andy Davidson -
Jim Reid -
Nigel Titley