Dear all, As we see ISPs and community would like to have more IPv4 space in use. I would like to ask a question what do people think about other side of IPv4 numeration space. Because we have in IPv4 a lot of addresses not in use at all but that space could be easy used. 240.0.0.0/4 Reserved (former Class E network) RFC 1700 it's 16 */8 networks. More then 256 Millions of routable and never used IPv4. 185/8 network has about 6.4M free and total RIPE has about 15M free IPv4 and we all say 185/8 will be enough for 2-3 years and rest - for some more time. But 256 M Ipv4 space could be enough for years! Space reserved for future Use. But will the future come to us or not? https://en.wikipedia.org/wiki/IPv4 https://tools.ietf.org/html/rfc1700 Is far as I see routers could easy start to use that IP space. People spend a lot of time and money to get some IPs but not to ask IANA to allow use this space. Technically it's very easy to start use IPs from such ranges. What does community thinks about it? Yuri
Implementation detail: many operating systems hard-code that range as invalid network space. The effort to make it available would be _less_ than getting everyone else in the world upgraded to IPv6. On 2016 Jun 11 (Sat) at 21:45:03 +0300 (+0300), NTX NOC wrote: :Dear all, : :As we see ISPs and community would like to have more IPv4 space in use. : :I would like to ask a question what do people think about other side of :IPv4 numeration space. Because we have in IPv4 a lot of addresses not in :use at all but that space could be easy used. : :240.0.0.0/4 Reserved (former Class E network) RFC 1700 : :it's 16 */8 networks. More then 256 Millions of routable and never used :IPv4. 185/8 network has about 6.4M free and total RIPE has about 15M :free IPv4 and we all say 185/8 will be enough for 2-3 years and rest - :for some more time. But 256 M Ipv4 space could be enough for years! : :Space reserved for future Use. But will the future come to us or not? : :https://en.wikipedia.org/wiki/IPv4 :https://tools.ietf.org/html/rfc1700 : :Is far as I see routers could easy start to use that IP space. People :spend a lot of time and money to get some IPs but not to ask IANA to :allow use this space. Technically it's very easy to start use IPs from :such ranges. : :What does community thinks about it? : :Yuri : -- Isn't it interesting that the same people who laugh at science fiction listen to weather forecasts and economists? -- Kelvin Throop III
On 11.06.2016 21:56, Peter Hessler wrote:
many operating systems hard-code that range as invalid network space.
Could you give any OS examples? I looks to my Juniper docs and see http://www.juniper.net/documentation/en_US/junos13.3/topics/topic-map/martia... It's not allowed by default but in one click you can make it work 240.0.0.0/4 orlonger -- allowed About IPv6 - still now in Russia there are no Home ISPs who gives IPv6 by default to customers. Nobody wants it, nobody needs it. Yuri
On Sat, 11 Jun 2016, NTX NOC wrote:
About IPv6 - still now in Russia there are no Home ISPs who gives IPv6 by default to customers. Nobody wants it, nobody needs it.
... and yet here you are, asking for obscure IPv4 blocks that are non-working in most operating systems available today, whilst all of them support IPv6 already. So "need" is in the eye of te beholder it seems. -- Mikael Abrahamsson email: swmike@swm.pp.se
Agree with Mikael. This is not a provocative question, but are NTX so worried about this proposal because it would affect their businesses selling and leasing IPv4 space? Their webpage mentions it at the top. The IP addresses aren't here to sell them. They're to be routed and used. This is just a tool. The IPv6 deployment is the only solution. On 06/11/16 22:27, Mikael Abrahamsson wrote:
On Sat, 11 Jun 2016, NTX NOC wrote:
About IPv6 - still now in Russia there are no Home ISPs who gives IPv6 by default to customers. Nobody wants it, nobody needs it.
... and yet here you are, asking for obscure IPv4 blocks that are non-working in most operating systems available today, whilst all of them support IPv6 already.
So "need" is in the eye of te beholder it seems.
-- Kind regards, CTO at *Foton Telecom CJSC* Tel.: +7 (499) 679-99-99 AS42861 on PeeringDB <http://as42861.peeringdb.com/>, Qrator <https://radar.qrator.net/as42861>, BGP.HE.NET <http://bgp.he.net/AS42861> http://ipv6actnow.org/
Not correct. My opinion is that all IPs space should be completely free for all members. It's like letters in the alphabet. You should not pay for letters, you should not pay for your unique name+surname (symbols that allow to identify you like IP address numbers). So to allow progress to come in we need to use abilities, that we have, reasonably. And here I asked about reserved IPv4 space. I am here in this discussions because we try to help people to get IP space easy and faster. And I am show here that there a lot of space that could be used. Yuri On 11.06.2016 22:33, Sergey wrote:
Agree with Mikael. This is not a provocative question, but are NTX so worried about this proposal because it would affect their businesses selling and leasing IPv4 space? Their webpage mentions it at the top.
The IP addresses aren't here to sell them. They're to be routed and used. This is just a tool. The IPv6 deployment is the only solution.
When a resource is scarce, either you define a high-enough price (GSM frequencies for instance) or you enforce a policy on how it's used. The issue is that even if you free 2 or 3 /8 from 240/4 or DoD, that's still a scarce resource, given the current growth and the forthcoming IoT invasion. David Ponzone
Le 13 juin 2016 à 19:09, NTX NOC <noc@ntx.ru> a écrit :
Not correct. My opinion is that all IPs space should be completely free for all members. It's like letters in the alphabet. You should not pay for letters, you should not pay for your unique name+surname (symbols that allow to identify you like IP address numbers).
So to allow progress to come in we need to use abilities, that we have, reasonably. And here I asked about reserved IPv4 space.
I am here in this discussions because we try to help people to get IP space easy and faster. And I am show here that there a lot of space that could be used.
Yuri
On 11.06.2016 22:33, Sergey wrote: Agree with Mikael. This is not a provocative question, but are NTX so worried about this proposal because it would affect their businesses selling and leasing IPv4 space? Their webpage mentions it at the top.
The IP addresses aren't here to sell them. They're to be routed and used. This is just a tool. The IPv6 deployment is the only solution.
*********************************************************************************************************** Le service MailSecure d'IPeva confirme l'absence de virus et de spam dans ce message. ***********************************************************************************************************
Hi, On Mon, Jun 13, 2016 at 08:09:50PM +0300, NTX NOC wrote:
And I am show here that there a lot of space that could be used.
You did not listen to the answers that stated that it's - *not* "a lot of" space - and "it can *not* be used" Gert Doering -- NetMaster -- have you enabled IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (0)89/32356-444 USt-IdNr.: DE813185279
On Mon, 13 Jun 2016, NTX NOC wrote:
Not correct. My opinion is that all IPs space should be completely free for all members. It's like letters in the alphabet. You should not pay for letters, you should not pay for your unique name+surname (symbols that allow to identify you like IP address numbers).
So to allow progress to come in we need to use abilities, that we have, reasonably. And here I asked about reserved IPv4 space.
I am here in this discussions because we try to help people to get IP space easy and faster. And I am show here that there a lot of space that could be used.
Well, using 240/3 isn't something that realistic. It is a lot easier to deply IPv6 than to get 240/3 working for any significant amount of users. We have run out of "letters" to use. The answer to the problem with "we've run out of letters" is to deploy IPv6. It's unfortunate that ISPs in your market aren't interested in Ipv6 deployment, but it's the only answer to your question. -- Mikael Abrahamsson email: swmike@swm.pp.se
Well, using 240/3 isn't something that realistic. It is a lot easier to deply IPv6 than to get 240/3 working for any significant amount of users.
Some may prefer easier ways (which is not that much easy to others) and some may not, My question is that is this working group the right place to discuss about the 240/3 or it should be done in higher level like between RIRs or IANA? Regards, Arash
Arash Naderpour wrote : My question is that is this working group the right place to discuss about the 240/3 or it should be done in higher level like between RIRs or IANA?
It has been done at higher levels multiple times in the last 15 years, and you are wasting your time and everyone else's. Broken record syndrome. Guys and girls, stop feeding the troll. Michel.
On Tue, 14 Jun 2016, Arash Naderpour wrote:
My question is that is this working group the right place to discuss about the 240/3 or it should be done in higher level like between RIRs or IANA?
You need to bring it to the IETF. IANA would most likely do what IETF asks it to do, and then the RIRs would follow suit. As was stated before (check out the link to the IETF draft, I'd imagine if you follow that trail you'll find discussions as well), the IETF took a look at this in 2008 or so, and it was deemed to be not practically feasible way of solving the problem. I'd say this hasn't changed. So most likely, if you go there and pitch the idea, you'll get the same reaction as you did here. -- Mikael Abrahamsson email: swmike@swm.pp.se
Good morning Arash, * "Arash Naderpour" <arash_mpc@parsun.com>
My question is that is this working group the right place to discuss about the 240/3 or it should be done in higher level like between RIRs or IANA?
RIPE AP-WG is not the right place to begin this process, the IETF is. The process would go something like this: You submit a draft to the IETF to direct IANA to do something with with 240/3, e.g., reclassify it as regular unicast IPv4 address space that may be distributed to the RIRs. You'll then need to gain consensus for your draft and have it published as an RFC. The /3 would then within six months be split up into five equal parts and be distributed to each RIR over a period of a few years. ~6.4 /8s per RIR, that is. The initial and biggest IANA->RIR trance would happen no later than six months after your RFC was published. (If you're not happy with that you'd need to seek global consensus between the five RIR communities to change the «Global Policy for Post Exhaustion IPv4 Allocation Mechanisms by the IANA» policy.) The RIPE NCC would add any address space received from the IANA in this manner to the so-called «last /8» pool. So assuming you've already received your final /22 under the current policy but want one or more additional allocations from 240/3, you'll at this point need to return to the RIPE AP-WG with a proposal to change the so-called «last /8» policy into something else that would facilitate that. Assuming you manage all of the above, all that remains in order to make 240/3 usable on the public Internet is to convince all the operating system/device/router vendors in the world to develop and release software/firmware updates to make 240/3 usable, and then of course to convince every network operator and end-user on the Internet to download and install these patches. Devices/software no longer being supported by the manufacturer would probably need to be replaced outright. If by some miracle you would be able to pull it all off, keep in mind that the ~107M addresses gained by the RIPE NCC would all be used up within two years if we return to the pre-depletion allocation policy and consumption rate. Ask yourself: «then what?» Maybe you can now see why folks are telling you that this would be a colossal waste of time and that your efforts would be much better spent on IPv6. With IPv6, the process is already underway and most of the above steps have already been completed, and at the end of that process we're actually covered for the rest of our lifetimes and beyond. Tore
But really, if you expect to be taken seriously, you should write your draft about 240/4, not 240/3 :) David Ponzone
Le 14 juin 2016 à 07:38, Tore Anderson <tore@fud.no> a écrit :
Good morning Arash,
* "Arash Naderpour" <arash_mpc@parsun.com>
My question is that is this working group the right place to discuss about the 240/3 or it should be done in higher level like between RIRs or IANA?
RIPE AP-WG is not the right place to begin this process, the IETF is.
The process would go something like this:
You submit a draft to the IETF to direct IANA to do something with with 240/3, e.g., reclassify it as regular unicast IPv4 address space that may be distributed to the RIRs. You'll then need to gain consensus for your draft and have it published as an RFC.
The /3 would then within six months be split up into five equal parts and be distributed to each RIR over a period of a few years. ~6.4 /8s per RIR, that is. The initial and biggest IANA->RIR trance would happen no later than six months after your RFC was published. (If you're not happy with that you'd need to seek global consensus between the five RIR communities to change the «Global Policy for Post Exhaustion IPv4 Allocation Mechanisms by the IANA» policy.)
The RIPE NCC would add any address space received from the IANA in this manner to the so-called «last /8» pool. So assuming you've already received your final /22 under the current policy but want one or more additional allocations from 240/3, you'll at this point need to return to the RIPE AP-WG with a proposal to change the so-called «last /8» policy into something else that would facilitate that.
Assuming you manage all of the above, all that remains in order to make 240/3 usable on the public Internet is to convince all the operating system/device/router vendors in the world to develop and release software/firmware updates to make 240/3 usable, and then of course to convince every network operator and end-user on the Internet to download and install these patches. Devices/software no longer being supported by the manufacturer would probably need to be replaced outright.
If by some miracle you would be able to pull it all off, keep in mind that the ~107M addresses gained by the RIPE NCC would all be used up within two years if we return to the pre-depletion allocation policy and consumption rate. Ask yourself: «then what?»
Maybe you can now see why folks are telling you that this would be a colossal waste of time and that your efforts would be much better spent on IPv6. With IPv6, the process is already underway and most of the above steps have already been completed, and at the end of that process we're actually covered for the rest of our lifetimes and beyond.
Tore
*********************************************************************************************************** Le service MailSecure d'IPeva confirme l'absence de virus et de spam dans ce message. ***********************************************************************************************************
On Tue, 14 Jun 2016, Tore Anderson wrote:
The /3 would then within six months be split up into five equal parts and be distributed to each RIR over a period of a few years. ~6.4 /8s
Well, it's only 16 /8s, so 3.2 /8s per RIR. Tore, great email summing up the problems with this proposal. -- Mikael Abrahamsson email: swmike@swm.pp.se
* Mikael Abrahamsson <swmike@swm.pp.se>
On Tue, 14 Jun 2016, Tore Anderson wrote:
The /3 would then within six months be split up into five equal parts and be distributed to each RIR over a period of a few years. ~6.4 /8s
Well, it's only 16 /8s, so 3.2 /8s per RIR.
Nnngh. Thanks David and Mikael. (This is why you should never write e-mails while not under the influence of coffee, kids.) Anyway this part needs corrections (emphasised with **) too:
If by some miracle you would be able to pull it all off, keep in mind that the **~54M** addresses gained by the RIPE NCC would all be used up within **one year** if we return to the pre-depletion allocation policy and consumption rate. Ask yourself: «then what?»
I'm assuming here an allocation rate of ~0.3 /8s per month, cf. https://labs.ripe.net/Members/wilhelm/global-patterns-in-ipv4-allocation-sta... Probably this estimate is way too low though, due to the unmet demand that has been building up in the LIRs over the course of the last four years. My guess is that we'd easily manage to fully deplete the first 240/4 IANA->RIR tranche (containing a /7) before the six months have passed before the second tranche (containing a /8) comes, and so on until the IANA recovered IPv4 pool is all gone. Tore
Hi, On Tue, Jun 14, 2016 at 11:55:12AM +1000, Arash Naderpour wrote:
Well, using 240/3 isn't something that realistic. It is a lot easier to deply IPv6 than to get 240/3 working for any significant amount of users.
Some may prefer easier ways (which is not that much easy to others) and some may not,
240/3 is not going to be easy. *Every* device out there would need to be changed (or at least *checked*) to ensure that it understands that these addresse are not special and can be used as normal unicast space. I could imagine that LI equipment that does not handle IPv6 will not handle Class E space either...
My question is that is this working group the right place to discuss about the 240/3 or it should be done in higher level like between RIRs or IANA?
It has been said before that 240/3 needs to be designated as unicast address space in the IETF first, then IANA could distribute to the RIRs. Pointers to the relevant drafts have been given. Gert Doering -- NetMaster -- have you enabled IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (0)89/32356-444 USt-IdNr.: DE813185279
This simply is not true: https://version6.ru/isp Also during the last ENOG in Moscow I've seen a lot of End-User ISPs actually trying to implement IPv6 right now. Making efforts to keep up legacy protocols running on the Internet is a step back. On 06/11/16 22:14, NTX NOC wrote:
About IPv6 - still now in Russia there are no Home ISPs who gives IPv6 by default to customers. Nobody wants it, nobody needs it. Yuri
-- Kind regards, CTO at *Foton Telecom CJSC* Tel.: +7 (499) 679-99-99 AS42861 on PeeringDB <http://as42861.peeringdb.com/>, Qrator <https://radar.qrator.net/as42861>, BGP.HE.NET <http://bgp.he.net/AS42861> http://ipv6actnow.org/
Hi, On Sat, Jun 11, 2016 at 10:14:11PM +0300, NTX NOC wrote:
About IPv6 - still now in Russia there are no Home ISPs who gives IPv6 by default to customers. Nobody wants it, nobody needs it.
The "nobody needs it" is a misconception. Direct your energy there to make people understand that IPv4 is game over. Gert Doering -- NetMaster -- have you enabled IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (0)89/32356-444 USt-IdNr.: DE813185279
Yuri, I wouldn't have a difficulty with it :) ... I dont see a reason why it wouldnt work ... although you would want everyone who is filtering bogons manually from their routers and the 240.0.0.0/4 has been considered a bogon for quite some time... so alot of people who do rudimentary prefix filtering on their border routers would have to update to make that range usable ... I have heard arguments that some Operating systems have that range filtered out and is non configurable... but I doubt the IPv6 adoption advocates ... or the IPv4 Sellers would like that idea too much as it would cause the price per ipv4 to collapse... I reckon this question has been asked before and I'm sure someone will point us to that discussion before... Hope this helps, On Sat, Jun 11, 2016 at 7:45 PM, NTX NOC <noc@ntx.ru> wrote:
Dear all,
As we see ISPs and community would like to have more IPv4 space in use.
I would like to ask a question what do people think about other side of IPv4 numeration space. Because we have in IPv4 a lot of addresses not in use at all but that space could be easy used.
240.0.0.0/4 Reserved (former Class E network) RFC 1700
it's 16 */8 networks. More then 256 Millions of routable and never used IPv4. 185/8 network has about 6.4M free and total RIPE has about 15M free IPv4 and we all say 185/8 will be enough for 2-3 years and rest - for some more time. But 256 M Ipv4 space could be enough for years!
Space reserved for future Use. But will the future come to us or not?
https://en.wikipedia.org/wiki/IPv4 https://tools.ietf.org/html/rfc1700
Is far as I see routers could easy start to use that IP space. People spend a lot of time and money to get some IPs but not to ask IANA to allow use this space. Technically it's very easy to start use IPs from such ranges.
What does community thinks about it?
Yuri
-- Kindest regards, Tom Smyth Mobile: +353 87 6193172 --------------------------------- PLEASE CONSIDER THE ENVIRONMENT BEFORE YOU PRINT THIS E-MAIL This email contains information which may be confidential or privileged. The information is intended solely for the use of the individual or entity named above. If you are not the intended recipient, be aware that any disclosure, copying, distribution or use of the contents of this information is prohibited. If you have received this electronic transmission in error, please notify me by telephone or by electronic mail immediately. Any opinions expressed are those of the author, not the company's .This email does not constitute either offer or acceptance of any contractually binding agreement. Such offer or acceptance must be communicated in writing. You are requested to carry out your own virus check before opening any attachment. Thomas Smyth accepts no liability for any loss or damage which may be caused by malicious software or attachments.
On Sat, 11 Jun 2016, NTX NOC wrote:
What does community thinks about it?
People have looked into this before. It's not feasible, not enough client OSes support it. People even tried this in the IETF, lots of years ago: https://tools.ietf.org/html/draft-savolainen-indicating-240-addresses-01 IPv4 stone is blead dry. Even if we doubled number of IPv4 addresses by means of some unknown magic, it wouldn't buy is any significant amount of time. The solution is IPv6. There is no other way to fix this. Direct your energy in that direction. -- Mikael Abrahamsson email: swmike@swm.pp.se
Hi,
As we see ISPs and community would like to have more IPv4 space in use.
I would like to ask a question what do people think about other side of IPv4 numeration space. Because we have in IPv4 a lot of addresses not in use at all but that space could be easy used.
240.0.0.0/4 Reserved (former Class E network) RFC 1700
I remember people looking into that years ago, and the conclusion was that in too many routers and operating systems the 240/4 block was hard-coded as unusable. I just checked the Linux source code and that restriction was removed there around 2008, but similar code was present in so many different places that it wasn't a viable solution. Remember that it wouldn't just be the organisation getting a block from 240/4, it would also affect everybody trying to communicate with them. Operating systems refusing to connect to a 240/4 address would make any website hosted on a 240/4 address badly reachable. Same for DNS servers hosted on such an address etc.
it's 16 */8 networks. More then 256 Millions of routable and never used IPv4.
That is actually not that much. In 2012 when we ran out of free IPv4 space for normal use the rate of allocation world-wide was more than a /8 per month. Even if we could use 240/4 these days, it would probably last us a year or so, and then we would be back where we are today. So in short: - 240/4 use is problematic - software needs to be changed in many places to make it usable - same for configurations (bogon filters etc) - it wouldn't last us much longer than a year anyway - we still need to move to IPv6 because we will have again run out Even shorter: we would use up 240/4 in less time than we would need to make it actually usable, so let's not. If we are changing stuff let's just spend our energy on implementing IPv6 instead. Cheers, Sander
Hi, On Sat, Jun 11, 2016 at 09:45:03PM +0300, NTX NOC wrote:
240.0.0.0/4 Reserved (former Class E network) RFC 1700
Many OSes do not support it, because it is not unicast address space (coming behind the multicast range 224.0.0.0/4). Some do, but if you put your webservers in there, you want to make sure that *all* clients can use it. So lots of software and hardware upgrades would be needed to ensure it's actually useful. And if you do that, why not do IPv6? The other reason why this is not a useful avenue to purchase is that 16 /8s would be burnt up in like 1.5 years time if we return to the old regime of "show me your need, and I give you enough address space for it" - many large providers could easily use up a few /12 here and there... bang, first /8 gone. Gert Doering -- NetMaster -- have you enabled IPv6 on something today...? SpaceNet AG Vorstand: Sebastian v. Bomhard Joseph-Dollinger-Bogen 14 Aufsichtsratsvors.: A. Grundner-Culemann D-80807 Muenchen HRB: 136055 (AG Muenchen) Tel: +49 (0)89/32356-444 USt-IdNr.: DE813185279
NTX NOC wrote: [...]
I would like to ask a question what do people think about other side of IPv4 numeration space. Because we have in IPv4 a lot of addresses not in use at all but that space could be easy used.
240.0.0.0/4 Reserved (former Class E network) RFC 1700
it's 16 */8 networks. More then 256 Millions of routable and never used IPv4. 185/8 network has about 6.4M free and total RIPE has about 15M free IPv4 and we all say 185/8 will be enough for 2-3 years and rest - for some more time. But 256 M Ipv4 space could be enough for years!
Space reserved for future Use. But will the future come to us or not?
https://en.wikipedia.org/wiki/IPv4 https://tools.ietf.org/html/rfc1700
Is far as I see routers could easy start to use that IP space. People spend a lot of time and money to get some IPs but not to ask IANA to allow use this space. Technically it's very easy to start use IPs from such ranges.
What does community thinks about it?
There are two I-Ds on this specific issue from 2008: https://tools.ietf.org/html/draft-fuller-240space-02 https://tools.ietf.org/html/draft-wilson-class-e-02 I believe progress on them stopped because the cost/benefit analysis meant it made much more sense to deploy IPv6. On the broader issue, decisions on what is and is not unicast IPv4 address space are made in the IETF. If you do want to head down this road, I suggest that you do so by writing an I-D and getting discussion on it in the IETF. The RFC Editor has a page, with appropriate links, on the process, here: https://www.rfc-editor.org/pubprocess/ Kind regards, Leo Vegoda
participants (12)
-
Arash Naderpour -
David Ponzone -
Gert Doering -
Leo Vegoda -
Michel Py -
Mikael Abrahamsson -
NTX NOC -
Peter Hessler -
Sander Steffann -
Sergey -
Tom Smyth -
Tore Anderson