Post-Exhaustion-phase "punishment" for early adopters
Hi, NANOG. Something's just struck me: every IPv4 allocation over a certain threshold has a monetary cost (sometimes in the tens of thousands of USD) and according to our RIR, the first equivalent IPv6 allocation is given as a freebie (to encourage migration). (Disclaimer: I'm on the Dark Continent of Africa) So once the "early" adopters migrate their networks to IPv6, there is no business need to maintain the IPv4 allocation and that will be returned to the free pool, since Business would see it as an unnecessary cost. This would seem to counteract the forced move to IPv6, since, once the early adopters move their services exclusively to IPv6 (or maintaining very small IPv4 blocks), there would be plenty of IPv4 space for the late adopters to request (after the RIR quarantine period, etc). Naturally, 6to4 functionality must remain for a while to interoperability reasons, so their resources would be available to the IPv6 world for time to come. Unless I'm misunderstanding the RIRs policy regarding IPv4 allocations; has it been stated by all RIRs that IPv4 blocks are unallocatable once the exhaustion phase kicks in? Or is there another mechanism to ensure that we don't hand out the space being handed back once IPv6 is the norm? :) Regards, -H.
On Feb 4, 2011, at 6:47 PM, Heinrich Strauss wrote:
So once the "early" adopters migrate their networks to IPv6, there is no business need to maintain the IPv4 allocation and that will be returned to the free pool, since Business would see it as an unnecessary cost.
Interesting reasoning. I would think that until we have pretty wide IPv4 implementation, the business need to keep the allocation is to talk with the people who have not yet implemented it. From a Reductio ad Absurdum perspective, imagine that facebook or youtube, now that they have implemented IPv6, felt obliged to give up their IPv4 allocation immediately? It would mean that they were out of business, which I should think might be an excellent business reason to not deploy IPv6.
On Feb 4, 2011, at 1:08 PM, Fred Baker wrote:
On Feb 4, 2011, at 6:47 PM, Heinrich Strauss wrote:
So once the "early" adopters migrate their networks to IPv6, there is no business need to maintain the IPv4 allocation and that will be returned to the free pool, since Business would see it as an unnecessary cost.
Interesting reasoning. I would think that until we have pretty wide IPv4 implementation, the business need to keep the allocation is to talk with the people who have not yet implemented it. From a Reductio ad Absurdum perspective, imagine that facebook or youtube, now that they have implemented IPv6, felt obliged to give up their IPv4 allocation immediately? It would mean that they were out of business, which I should think might be an excellent business reason to not deploy IPv6.
Exactly. Which means that folks deploying IPv6 will keep their IPv4 until no longer needed. Which in turn means that the value of redeploying those returned addresses would be minimal - the Internet would be dominantly IPv6 at that point in time. Cheers, -Benson
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Feb 4, 2011, at 10:47 AM, Heinrich Strauss wrote:
So once the "early" adopters migrate their networks to IPv6, there is no business need to maintain the IPv4 allocation and that will be returned to the free pool, since Business would see it as an unnecessary cost.
However, doing enough of a migration to be able to actually free up the IPv4 addresses, in advance of the devices using them dying of old age or being naturally cycled out, also imposes a cost, and likely a high one relative to the RIR maintenance fees, so it's my guess that the rate of IPv4 returns may not be too fast.
This would seem to counteract the forced move to IPv6, since, once the early adopters move their services exclusively to IPv6 (or maintaining very small IPv4 blocks), there would be plenty of IPv4 space for the late adopters to request.
If you look at the actual numbers involved, I think you'll find that the "plenty" would actually be quite small and in quite small chunks, relative to what the industry could use, if they weren't trying to get over to v6. So I doubt it will have that much effect.
Has it been stated by all RIRs that IPv4 blocks are unallocatable once the exhaustion phase kicks in? Or is there another mechanism to ensure that we don't hand out the space being handed back once IPv6 is the norm?
No, and in fact, I believe all the RIRs will probably do a reasonably brisk business in reclamation and reallocation, albeit in ever smaller blocks. One way to think about it is that there's no particular reason to think that the rate of increase in the number of IPv4 prefixes in the global BGP routing table will slack off, therefore those prefixes will each simply be smaller and smaller, over time. More or less. Speaking not particularly with my ARIN-board-hat-on, -Bill -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (Darwin) iEYEARECAAYFAk1MT0YACgkQGvQy4xTRsBH3RwCgpfo6JbOC8aYnsCu3h5B9++6u qA4AoJVds96Ua9dvqx2RDUw9qBZcyOQK =aMk8 -----END PGP SIGNATURE-----
On Feb 4, 2011, at 1:11 PM, Bill Woodcock wrote:
No, and in fact, I believe all the RIRs will probably do a reasonably brisk business in reclamation and reallocation, albeit in ever smaller blocks.
As holder of a small block, this scares and irritates me. It scares me that I might lose my autonomy and future expansion through no fault of my own, and it irritates me that the reason I may be forced to give up my address space will probably be to satisfy the internet's desperate need for more spam cannons.
On Feb 4, 2011, at 3:39 PM, Daniel Seagraves wrote:
On Feb 4, 2011, at 1:11 PM, Bill Woodcock wrote:
No, and in fact, I believe all the RIRs will probably do a reasonably brisk business in reclamation and reallocation, albeit in ever smaller blocks.
As holder of a small block, this scares and irritates me. It scares me that I might lose my autonomy and future expansion through no fault of my own, and it irritates me that the reason I may be forced to give up my address space will probably be to satisfy the internet's desperate need for more spam cannons.
If you are using your block, why would you worry? If not are not using your block, why would you need it? -- TTFN, patrick
On Feb 4, 2011, at 3:41 PM, Patrick W. Gilmore wrote:
On Feb 4, 2011, at 3:39 PM, Daniel Seagraves wrote:
On Feb 4, 2011, at 1:11 PM, Bill Woodcock wrote:
No, and in fact, I believe all the RIRs will probably do a reasonably brisk business in reclamation and reallocation, albeit in ever smaller blocks.
As holder of a small block, this scares and irritates me. It scares me that I might lose my autonomy and future expansion through no fault of my own, and it irritates me that the reason I may be forced to give up my address space will probably be to satisfy the internet's desperate need for more spam cannons.
If you are using your block, why would you worry?
If not are not using your block, why would you need it?
Likely because some devices still don't implement IPv6 bootstrap. Try to recover a Cisco router via TFTP boot in an IPv6 only environment. I have been trying to remind my vendors to think about IPv6 first and IPv4 as a secondary capability to supplement it, I do encourage everyone to make this part of your procurement of any equipment in 2011 and beyond. eg: If your DNS provider doesn't do IPv6, switch. (has tucows solved the AAAA glue issue yet? i think i need to switch... and no, i don't feel like using a hack process via a web form, I actually want real automated interfaces and support...) - Jared
On Feb 4, 2011, at 4:45 PM, Jared Mauch wrote:
On Feb 4, 2011, at 3:41 PM, Patrick W. Gilmore wrote:
On Feb 4, 2011, at 3:39 PM, Daniel Seagraves wrote:
On Feb 4, 2011, at 1:11 PM, Bill Woodcock wrote:
No, and in fact, I believe all the RIRs will probably do a reasonably brisk business in reclamation and reallocation, albeit in ever smaller blocks.
As holder of a small block, this scares and irritates me. It scares me that I might lose my autonomy and future expansion through no fault of my own, and it irritates me that the reason I may be forced to give up my address space will probably be to satisfy the internet's desperate need for more spam cannons.
If you are using your block, why would you worry?
If not are not using your block, why would you need it?
Likely because some devices still don't implement IPv6 bootstrap. Try to recover a Cisco router via TFTP boot in an IPv6 only environment.
I have been trying to remind my vendors to think about IPv6 first and IPv4 as a secondary capability to supplement it, I do encourage everyone to make this part of your procurement of any equipment in 2011 and beyond.
eg: If your DNS provider doesn't do IPv6, switch. (has tucows solved the AAAA glue issue yet? i think i need to switch... and no, i don't feel like using a hack process via a web form, I actually want real automated interfaces and support...)
I'm a little confused. Sounds like the things you are talking about all fall into the "if you are using your block" category, so he shouldn't worry. ARIN should not reclaim a block that is in use. Unless I am confused? (Happens a lot, especially as I get older.) -- TTFN, patrick
On Feb 4, 2011, at 3:51 PM, Patrick W. Gilmore wrote:
I'm a little confused. Sounds like the things you are talking about all fall into the "if you are using your block" category, so he shouldn't worry.
ARIN should not reclaim a block that is in use. Unless I am confused? (Happens a lot, especially as I get older.)
How many addresses do I have to be using for it to count as in use? How high will that number go in the next few months/years? We have a very old /24 direct allocation from the stone age, when we were a dialup ISP. The company still exists, we just aren't providing dialup service anymore. We still have a couple of our web-hosting customers, but for the most part we've moved on to running an unrelated web-based service. Having our own address space is nice because it means we don't have to worry about stepping on anyone's AUP, we can go multi-homed later as the usage goes up, and we don't have to worry about running out of space as the web service grows. The problem is that while we met the eligibility requirements for an ipv4 direct allocation back when we got it, the requirements have changed over time and we no longer meet the eligibility requirements for an ipv4 direct allocation. (We've shrunk quite a bit) As demand for the remaining ipv4 addresses goes up, ARIN might decide that since we're ineligible for an allocation under the current rules, we're no longer eligible to maintain the space we have, and take it away from us. As the remaining space gets smaller, I expect that the number needed to justify keeping my addresses is going to go up. I fear I'm already on thin ice.
On Fri, Feb 4, 2011 at 4:28 PM, Daniel Seagraves <dseagrav@humancapitaldev.com> wrote:
On Feb 4, 2011, at 3:51 PM, Patrick W. Gilmore wrote:
How many addresses do I have to be using for it to count as in use? How high will that number go in the next few months/years?
The most important thing to ensure "usage" is recognized is that the entire address space is announced plus routed, at least one valid host exists, reverse DNS servers are live, AND ensure that all contact details in the WHOIS database are present, accurate,and up to date; including phone numbers, postal address, and e-mails for the organization contacts. You might look into the option of signing the Legacy RSA: https://www.arin.net/resources/legacy/ Available until Dec 2011; If your allocation predated ARIN. I doubt the community is going to scour through all the /24 allocations and try to reclaim them, however. It's not that legacy /24 allocations don't matter, if they are entirely derelict, but the /8s are the ones that "sort of" matter... sort of, because a /8 reclaimed could provide a few months of IP addresses for a RIR. If you have a RSA, you are not really subject to the RIR taking IP addresses that you were already allocated (unless you had violated a policy, for example, by submitting a false application, or you no longer have a justified need for the IPs). Changes to the policy for new allocations doesn't mean existing allocations are cancelled, just because the new rule doesn't allow them. Probably it would not be too fair to try to answer a question that the community hasn't defined an answer for (yet at least) through any RIR policy, in regards to "how much usage" is usage. "Usage" is probably going to mean something like "globally" routed, at least one host exists in each /24, and the addresses are being used for the purpose originally in the application.. That would be what "used" means, but not what "efficiently utilized" means. It's not likely but conceivable, that the RIRs could decide to make a policy of reviewing legacy resources and revoking allocations with bad contact info, or apply an "efficient usage" criterion requiring return/renumbering, for legacy resource holders who have no RSA. -- -JH
----- Original Message -----
From: "Jimmy Hess" <mysidia@gmail.com>
On Fri, Feb 4, 2011 at 4:28 PM, Daniel Seagraves <dseagrav@humancapitaldev.com> wrote:
On Feb 4, 2011, at 3:51 PM, Patrick W. Gilmore wrote:
How many addresses do I have to be using for it to count as in use? How high will that number go in the next few months/years?
The most important thing to ensure "usage" is recognized is that the entire address space is announced plus routed, at least one valid host exists, reverse DNS servers are live, AND ensure that all contact details in the WHOIS database are present, accurate,and up to date; including phone numbers, postal address, and e-mails for the organization contacts.
You might look into the option of signing the Legacy RSA: https://www.arin.net/resources/legacy/ Available until Dec 2011; If your allocation predated ARIN.
I gather that if you're considering signing an LRSA, you might want to take *technically competent* legal advice first. Cheers, -- jra
-----Original Message----- From: Jimmy Hess [mailto:mysidia@gmail.com]
The most important thing to ensure "usage" is recognized is that the entire address space is announced plus routed,
I don't speak on behalf of a community, but in the past there have been people reminding the ARIN community that there are valid uses for address space, contemplated by rfc2050, in addition to routing on the public Internet.
You might look into the option of signing the Legacy RSA: https://www.arin.net/resources/legacy/ Available until Dec 2011; If your allocation predated ARIN.
Yes. You might decide you don't like some provision, but it would be careless not to look into it.
I doubt the community is going to scour through all the /24 allocations and try to reclaim them, however. It's not that legacy /24 allocations don't matter, if they are entirely derelict, but the /8s are the ones that "sort of" matter... sort of, because a /8 reclaimed could provide a few months of IP addresses for a RIR.
Agree; according to https://www.arin.net/knowledge/statistics/index.html ARIN has been issuing 20,000 - 50,000 /24 per month for the past few months. A /24 wouldn't extend runout time by a full minute.
It's not likely but conceivable, that the RIRs could decide to make a policy of reviewing legacy resources and revoking allocations with bad contact info, or apply an "efficient usage" criterion requiring return/renumbering, for legacy resource holders who have no RSA.
That would be an exciting debate. Lee
On Feb 4, 2011, at 2:28 PM, Daniel Seagraves wrote:
On Feb 4, 2011, at 3:51 PM, Patrick W. Gilmore wrote:
I'm a little confused. Sounds like the things you are talking about all fall into the "if you are using your block" category, so he shouldn't worry.
ARIN should not reclaim a block that is in use. Unless I am confused? (Happens a lot, especially as I get older.)
How many addresses do I have to be using for it to count as in use? How high will that number go in the next few months/years?
We have a very old /24 direct allocation from the stone age, when we were a dialup ISP. The company still exists, we just aren't providing dialup service anymore. We still have a couple of our web-hosting customers, but for the most part we've moved on to running an unrelated web-based service. Having our own address space is nice because it means we don't have to worry about stepping on anyone's AUP, we can go multi-homed later as the usage goes up, and we don't have to worry about running out of space as the web service grows. The problem is that while we met the eligibility requirements for an ipv4 direct allocation back when we got it, the requirements have changed over time and we no longer meet the eligibility requirements for an ipv4 direct allocation. (We've shrunk quite a bit) As demand for the remaining ipv4 addresses goes up, ARIN might decide that since we're ineligible for an allocation under the current rules, we're no longer eligible to maintain the space we have, and take it away from us.
As the remaining space gets smaller, I expect that the number needed to justify keeping my addresses is going to go up. I fear I'm already on thin ice.
If you don't sign the LRSA, who knows. If you sign the LRSA, you're completely protected regardless of future policy changes. Owen
Hi,
If you are using your block, why would you worry?
If not are not using your block, why would you need it?
You may define "using" Hint: even IPs not pingable from the Internet are being used. Not everyone is an ISP/Webhoster ... with public services. -- Viele Grüße / Kind Regards / Cordiali Saluti / Met vriendelijke groet Ralph J.Mayer xmpp:rmayer@vinotech.de www.vinoblog.de mailto:rmayer@vinotech.de
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Feb 4, 2011, at 1:11 PM, Bill Woodcock wrote:
No, and in fact, I believe all the RIRs will probably do a reasonably brisk business in reclamation and reallocation, albeit in ever smaller blocks.
On Feb 4, 2011, at 12:39 PM, Daniel Seagraves wrote:
As holder of a small block, this scares and irritates me. It scares me that I might lose my autonomy and future expansion through no fault of my own, and it irritates me that the reason I may be forced to give up my address space will probably be to satisfy the internet's desperate need for more spam cannons.
Excuse me, "reclamation" was probably the wrong word choice on my part. What I was intending to convey was "processing of returned blocks." And the use for the ever-smaller blocks is not for spammers, but for the IPv4 side of 4-to-6 NATs. -Bill -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (Darwin) iEYEARECAAYFAk1MZSwACgkQGvQy4xTRsBEdqACcDnngVari/dTZrt+ha9P8trct 7J4AoJDftyNiU/lB2+nHZPJrTlIkzJGE =Aaf9 -----END PGP SIGNATURE-----
On Feb 4, 2011, at 10:47 AM, Heinrich Strauss wrote:
Hi, NANOG.
Something's just struck me: every IPv4 allocation over a certain threshold has a monetary cost (sometimes in the tens of thousands of USD) and according to our RIR, the first equivalent IPv6 allocation is given as a freebie (to encourage migration). (Disclaimer: I'm on the Dark Continent of Africa)
So once the "early" adopters migrate their networks to IPv6, there is no business need to maintain the IPv4 allocation and that will be returned to the free pool, since Business would see it as an unnecessary cost.
This would seem to counteract the forced move to IPv6, since, once the early adopters move their services exclusively to IPv6 (or maintaining very small IPv4 blocks), there would be plenty of IPv4 space for the late adopters to request (after the RIR quarantine period, etc). Naturally, 6to4 functionality must remain for a while to interoperability reasons, so their resources would be available to the IPv6 world for time to come.
Unless I'm misunderstanding the RIRs policy regarding IPv4 allocations; has it been stated by all RIRs that IPv4 blocks are unallocatable once the exhaustion phase kicks in? Or is there another mechanism to ensure that we don't hand out the space being handed back once IPv6 is the norm? :)
Regards, -H.
The big providers will not be deprecating their IPv4 addresses until there is no longer a significant IPv4 internet. At that time, smaller providers could probably get them, but, they will need IPv6 in order to talk to most of the world. Owen
participants (12)
-
Benson Schliesser -
Bill Woodcock -
Daniel Seagraves -
Fred Baker -
Heinrich Strauss -
Jared Mauch -
Jay Ashworth -
Jimmy Hess -
Lee Howard -
Owen DeLong -
Patrick W. Gilmore -
Ralph J.Mayer