2011-02 New Policy Proposal (Removal of multihomed requirement for IPv6)
Dear Colleagues, A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion. You can find the full proposal at: http://www.ripe.net/ripe/policies/proposals/2011-02 We encourage you to review this proposal and send your comments to <address-policy-wg@ripe.net> before 13 May 2011. Regards Emilio Madaio Policy Development Officer RIPE NCC
* Emilio Madaio:
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
You can find the full proposal at:
I think this should be amended to clarify that PI space can be assigned to employees, customers etc. independent of device ownership, using services like RA, reflecting the previous discussion on this list. -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99
Hi,
I think this should be amended to clarify that PI space can be assigned to employees, customers etc. independent of device ownership, using services like RA, reflecting the previous discussion on this list.
To keep things manageable let's not combine different things in one proposal. I understand that there are many people who would like to 'no sub-assignments' part of the PI policy but that is a different issue than the multihoming requirement. Let's discuss that separately. Thanks, Sander APWG co-chair
On Fri, 15 Apr 2011, Emilio Madaio wrote:
We encourage you to review this proposal and send your comments to <address-policy-wg@ripe.net> before 13 May 2011.
In reading this proposal, I note that the "opposing"-section contains text that opposes the opposing argument. I also feel the opposing text is biased. Generally us ISPs are not worried about RIB capacity, we're worried about FIB capacity. Every PI network carries a global cost because of a DFZ slot that all DFZ routers need to carry. Nonetheless, the important factor for me is: Will it cost the same amount of money per year? This is "tragedy of the commons" problem, so as long as the cost for holding the PI resource is the same before and after this proposal, I don't have any problems with it. If it's all of a sudden a lot cheaper, I oppose the change. Rationale: If someone is willing to spend a substantial amount of money per year to hold a PI, they have a real business need and are willing to back this with cash because it's still cheaper for them. If they don't, then please stop using DFZ resources for no real need. The policy should reflect this fact. If it does after the proposed change, I'm all for it. -- Mikael Abrahamsson email: swmike@swm.pp.se
Hi Michael,
Nonetheless, the important factor for me is: Will it cost the same amount of money per year? This is "tragedy of the commons" problem, so as long as the cost for holding the PI resource is the same before and after this proposal, I don't have any problems with it.
If it's all of a sudden a lot cheaper, I oppose the change.
Initially the proposal did include a change in the cost for PI IPv6, however it was decided to not include that into this discussion as cost for PI isn't decided within the community but in the AGM meeting. Having said this, there will be a public discussion behind this to get input from the community to review the cost of specifically PI IPv6 as input for the chairs to take into the AGM meeting. It wasn't my intention to ask for cheaper PI IPv6, the opposite in fact, I personally think that PI IPv6 should be more expensive than currently especially when the multihoming requirement would be dropped. The proposal as it is, is to have the same requirements for PI IPv6 as one would face if they would signup to become a LIR. Kind regards, Erik Bais
Erik Bais wrote:
Hi Michael,
Nonetheless, the important factor for me is: Will it cost the same amount of money per year? This is "tragedy of the commons" problem, so as long as the cost for holding the PI resource is the same before and after this proposal, I don't have any problems with it.
If it's all of a sudden a lot cheaper, I oppose the change.
Initially the proposal did include a change in the cost for PI IPv6, however it was decided to not include that into this discussion as cost for PI isn't decided within the community but in the AGM meeting.
Having said this, there will be a public discussion behind this to get input from the community to review the cost of specifically PI IPv6 as input for the chairs to take into the AGM meeting.
It wasn't my intention to ask for cheaper PI IPv6, the opposite in fact, I personally think that PI IPv6 should be more expensive than currently especially when the multihoming requirement would be dropped.
We should keep in mind that the RIPE NCC, working (+ charging for its services) according to the policies as agreed by the Community is NOT an internet tax and fund-redistribution entity! Usually, the cost for a particular service is set to reflect the effort in the NCC to provide the registration (and associated) services. There is no way to redistribute a "FIB-entry-surcharge" to whatever or whoever believes to "suffer" from this proposed change in policy.
The proposal as it is, is to have the same requirements for PI IPv6 as one would face if they would signup to become a LIR.
Not really. It would only remove special (and imho artificial to begin with) *operational* requirments (i.e. multi-homing, ASing) from the access to the resource. Registering a prefix in the Resource Registry by the RIPE NCC should require the same effort as IPv4 PI and the current IPv6 PI. Actually - it should become cheaper - effortwise - because the NCC can drop the superficial and ineffective "checks" made after assignment, to assess if the MH requirements were or still are met.
Kind regards, Erik Bais
Wilfried.
* Wilfried Woeber:
We should keep in mind that the RIPE NCC, working (+ charging for its services) according to the policies as agreed by the Community is NOT an internet tax and fund-redistribution entity!
I agree wholeheartedly.
There is no way to redistribute a "FIB-entry-surcharge" to whatever or whoever believes to "suffer" from this proposed change in policy.
And more fine-grained routing choices often improve some desirable metrics globally, and not just locally at one site. In any case, update churn is reportedly the worse problem today. -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99
On Tue, 19 Apr 2011, Wilfried Woeber, UniVie/ACOnet wrote:
There is no way to redistribute a "FIB-entry-surcharge" to whatever or whoever believes to "suffer" from this proposed change in policy.
As I stated before, I don't care if they get the money in cash from the bank and BURN them at a BBQ, as long as the money is paid so as to stop people using it who don't really have any real business use for it. I don't want the money, I just want there to be *substantial cost* to take up a DFZ slot. Preferrably it should be paid per slot as well, so people de-aggregating their blocks have to pay more, but I don't know any way to do that. -- Mikael Abrahamsson email: swmike@swm.pp.se
* Mikael Abrahamsson:
I don't want the money, I just want there to be *substantial cost* to take up a DFZ slot.
And degrade overall routing quality because networks cannot afford announcing shorter paths? I don't think this is a good idea.
Preferrably it should be paid per slot as well, so people de-aggregating their blocks have to pay more, but I don't know any way to do that.
Currently, the policy documents spell out clearly that a prefix assignment is not a guarantuee that the prefix will show up in the DFZ. Neither the RIPE community nor the RIPE NCC control the DFZ, so I'm not sure how to change this, and getting some control seems to be a necessary first step before charging for DFZ slots (that is, make sure that you can actually deliver the product you're trying to sell). -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99
On Thu, 21 Apr 2011, Florian Weimer wrote:
* Mikael Abrahamsson:
I don't want the money, I just want there to be *substantial cost* to take up a DFZ slot.
And degrade overall routing quality because networks cannot afford announcing shorter paths? I don't think this is a good idea.
Please elaborate. I don't understand your reasoning. How does de-aggregation improve routing quality? -- Mikael Abrahamsson email: swmike@swm.pp.se
* Mikael Abrahamsson:
On Thu, 21 Apr 2011, Florian Weimer wrote:
* Mikael Abrahamsson:
I don't want the money, I just want there to be *substantial cost* to take up a DFZ slot.
And degrade overall routing quality because networks cannot afford announcing shorter paths? I don't think this is a good idea.
Please elaborate. I don't understand your reasoning. How does de-aggregation improve routing quality?
It can make it possible for your network to choose a better path to the destination, even according to your own set of cost metrics. -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99
On Thu, 21 Apr 2011, Florian Weimer wrote:
Please elaborate. I don't understand your reasoning. How does de-aggregation improve routing quality?
It can make it possible for your network to choose a better path to the destination, even according to your own set of cost metrics.
That is not elaborating. -- Mikael Abrahamsson email: swmike@swm.pp.se
On Tue, Apr 19, 2011 at 01:02:59PM +0000, Wilfried Woeber, UniVie/ACOnet wrote:
Actually - it should become cheaper - effortwise - because the NCC can drop the superficial and ineffective "checks" made after assignment, to assess if the MH requirements were or still are met.
And it doesn't have to consider "need" regarding the amount of space requested, if not more than /48. And even if more, the rules are far simpler than with IPv4 where each and every address has to be justified. So in fact, the cost for NCC to process+maintain IPv6 PI should be (significantly?) less than with IPv4 PI. Best regards, Daniel -- CLUE-RIPE -- Jabber: dr@cluenet.de -- dr@IRCnet -- PGP: 0xA85C8AA0
On 15-04-2011 12:01, Mikael Abrahamsson wrote:
If someone is willing to spend a substantial amount of money per year to hold a PI, they have a real business need and are willing to back this with cash because it's still cheaper for them. If they don't, then please stop using DFZ resources for no real need.
Not everything on the internet is a bussiness with needs and a large bag of cash. Please also consider the non-profit organisations with valid reasons of running their own PI network. -- Michiel Klaver IT Professional
On 15 Apr 2011, at 12:04, Michiel Klaver wrote:
Not everything on the internet is a bussiness with needs and a large bag of cash. Please also consider the non-profit organisations with valid reasons of running their own PI network.
If an organisation like that can't afford a few hundred euro for addresses and an ASN, it probably can't afford any money for servers or routers or connectivity or co-lo space or software or a system/ network admin, or lawyers to review contracts, etc, etc. These things are all unavoidable costs of doing business on the interweb, even for non-profits. Supporting (under-funded) public benefit organisations is all well and good. But is it fair to expect the NCC membership to pick up the bill? It doesn't seem right to push costs I can't or won't or don't like to pay on to someone else. Though it is of course lovely for me if I can manipulate the system to decide how your money gets spent in ways for my benefit. Nice work if you can get it...
Hallo Jim, du schrobst:
On 15 Apr 2011, at 12:04, Michiel Klaver wrote:
Not everything on the internet is a bussiness with needs and a large bag of cash. Please also consider the non-profit organisations with valid reasons of running their own PI network. If an organisation like that can't afford a few hundred euro for addresses and an ASN, it probably can't afford any money for servers or routers or connectivity or co-lo space or software or a system/ network admin, or lawyers to review contracts, etc, etc. These things are all unavoidable costs of doing business on the interweb, even for non-profits.
I am working for such a non profit organization voluntarely (as all of our staff does) and I can assure you that fundraising money for servers/routers is not easy for us. When it comes to periodic costs, it is always a trade off between more bandwith/more servers (= increased power costs) and other things (like address space, lir membership fees, etc). So while most end users of independet ressources might not even notice a shift from 50 euros per anno to 500 euros per anno, at least for us that would make things seriously more difficult. After all, thats why we decided not to become a LIR (eventhough it would otherwise fit better then usage of independend ressources).
Supporting (under-funded) public benefit organisations is all well and good. But is it fair to expect the NCC membership to pick up the bill?
What bill exactly? I don't see much costs generated by users of independent ressources. Regards, Immo
On 15 April 2011 13:22, Immo 'FaUl' Wehrenberg <immo.ripe@be.free.de> wrote:
I am working for such a non profit organization voluntarely (as all of our staff does) and I can assure you that fundraising money for servers/routers is not easy for us. When it comes to periodic costs, it is always a trade off between more bandwith/more servers (= increased power costs) and other things (like address space, lir membership fees, etc). So while most end users of independet ressources might not even notice a shift from 50 euros per anno to 500 euros per anno, at least for us that would make things seriously more difficult. After all, thats why we decided not to become a LIR (eventhough it would otherwise fit better then usage of independend ressources).
I take your point.. but I have to agree with Jim on this one. I have worked for various non-profit organisations and also understand the financial pressures one is under. However non-profit or not-for-profit does not mean, for all intents and purposes, that it should spend aim to spend more than it can afford. Regardless of whether an organisation is for-profit or non-profit, it is a business and must be run like one. Just look at the many charities out there that also run like businesses... and non-profit does not mean it can not make a profit.. just that it has to work on profit-loss neutrality.. and ensure it's costs are met. My biggest bugbear in some non-profits I have worked in was that the 'real' costs were swept under the carpet... Taking it to it's conclusion, if every volunteer in the organisation was to be paid and all costs were considered, what would be the real cost of running the organisation? If it has not been done, then in my opinion, it is a good exercise to do as it will give you the tru value of running the organisation so that in case donations do drop at least you can be prepared with savings.
Supporting (under-funded) public benefit organisations is all well and
good. But is it fair to expect the NCC membership to pick up the bill? What bill exactly? I don't see much costs generated by users of independent ressources.
You mention you are a volunteer in a non-profit organisation. I honestly salute you. However, if you were to charge for your time, how much would the organisation have to pay you? In the same way, whatever resource is being used via RIPE, is a cost to the NCC and that needs to be paid somehow. If it does not come from the PI crowd, then it will come from the membership. That is the 'bill' that Jim is referring to. Regards Denesh
On Fri, Apr 15, 2011 at 14:40, Denesh Bhabuta <db-ml@cyberstrider.net> wrote:
You mention you are a volunteer in a non-profit organisation. I honestly salute you. However, if you were to charge for your time, how much would the organisation have to pay you?
This is an excellent argument in favor of lowering costs for NFPs, not against it. But as this is a moral, not a factual, question, it's probably not too much use arguing the point.
In the same way, whatever resource is being used via RIPE, is a cost to the NCC and that needs to be paid somehow. If it does not come from the PI crowd, then it will come from the membership. That is the 'bill' that Jim is referring to.
Realistically speaking, there is not much in actual additional costs other than the time to support requests. This could or could not be reflected in the distribution and cut-off limits of the LIR sizing which, in turn, can be influenced and decided upon by the RIPE community. I am not saying that this should be done. I am merely stating that this could make sense and that it's not as clear-cut as you seem to make it appear. Richard
On Fri, 15 Apr 2011, Richard Hartmann wrote:
On Fri, Apr 15, 2011 at 14:40, Denesh Bhabuta <db-ml@cyberstrider.net> wrote:
You mention you are a volunteer in a non-profit organisation. I honestly salute you. However, if you were to charge for your time, how much would the organisation have to pay you?
This is an excellent argument in favor of lowering costs for NFPs, not against it.
I wonder whether your organization has approached the electric company in order to receive a NPO discount. -Hank
But as this is a moral, not a factual, question, it's probably not too much use arguing the point.
In the same way, whatever resource is being used via RIPE, is a cost to the NCC and that needs to be paid somehow. If it does not come from the PI crowd, then it will come from the membership. That is the 'bill' that Jim is referring to.
Realistically speaking, there is not much in actual additional costs other than the time to support requests.
This could or could not be reflected in the distribution and cut-off limits of the LIR sizing which, in turn, can be influenced and decided upon by the RIPE community.
I am not saying that this should be done. I am merely stating that this could make sense and that it's not as clear-cut as you seem to make it appear.
Richard
I think we should always have a level and equal pricing model for PI, I see no reason why NFP organisations should be subsidised by the same standard that organisations that we may find morally objectionable shouldn't be charged more. -----Original Message----- From: address-policy-wg-admin@ripe.net [mailto:address-policy-wg-admin@ripe.net] On Behalf Of Richard Hartmann Sent: 15 April 2011 13:58 To: Denesh Bhabuta Cc: Immo 'FaUl' Wehrenberg; address-policy-wg@ripe.net Subject: Re: [address-policy-wg] the cost of PI space On Fri, Apr 15, 2011 at 14:40, Denesh Bhabuta <db-ml@cyberstrider.net> wrote:
You mention you are a volunteer in a non-profit organisation. I honestly salute you. However, if you were to charge for your time, how much would the organisation have to pay you?
This is an excellent argument in favor of lowering costs for NFPs, not against it. But as this is a moral, not a factual, question, it's probably not too much use arguing the point.
In the same way, whatever resource is being used via RIPE, is a cost to the NCC and that needs to be paid somehow. If it does not come from the PI crowd, then it will come from the membership. That is the 'bill' that Jim is referring to.
Realistically speaking, there is not much in actual additional costs other than the time to support requests. This could or could not be reflected in the distribution and cut-off limits of the LIR sizing which, in turn, can be influenced and decided upon by the RIPE community. I am not saying that this should be done. I am merely stating that this could make sense and that it's not as clear-cut as you seem to make it appear. Richard
On 15 Apr 2011, at 13:58, Richard Hartmann wrote:
This is an excellent argument in favor of lowering costs for NFPs, not against it.
But as this is a moral, not a factual, question, it's probably not too much use arguing the point.
Indeed. However it's rather ironic that some non-profits who are not NCC members seem to be complaining about fees. Since they can't/won't become members they're not in a position to influence the charging scheme.
Richard, On Fri, Apr 15, 2011 at 8:58 AM, Richard Hartmann <richih.mailinglist@gmail.com> wrote:
In the same way, whatever resource is being used via RIPE, is a cost to the NCC and that needs to be paid somehow. If it does not come from the PI crowd, then it will come from the membership. That is the 'bill' that Jim is referring to.
Realistically speaking, there is not much in actual additional costs other than the time to support requests.
That is a quite strong statement. The RIPE NCC has an operational running cost. What the community gets is a database where you can apply for, register and unregister some Internet resources, and in between they keep them there. Today there is a recurring cost based on an averaged load of maintaining resources (NFP/FP makes zero distinction for RIPE NCC here), and a setup fee for LIRs. The recurring resource costs, for simplicity, also cover the application costs of said resources - a cost that with IPv6 has a real chance to go down, if the NCC is sincere in simplifying the processes and aligning them with reality. (I'm clear: the community sets policies and the NCC implements them in their hostmasters.) If your statement was actually true, please do explain the PI / ASN / LIR pricing model of today and where these costs goes, if "there is not much in actual additional costs" of giving out and maintaining PI resources at discounted prices to some. The real truth is that any subsidizing will be just that; subsidizing. I am in favor of the current simple PI / ASN cost model as it is and oppose subsidizing as it will lead to unbalance. Thus, I want any change in pricing to be discussed for the general PI resource, following today's model. Regards, Martin
Prefix: My statement was probably overly broad and, apparently, I didn't flesh it out enough to make my point. On Sun, Apr 17, 2011 at 04:03, Martin Millnert <millnert@gmail.com> wrote:
If your statement was actually true, please do explain the PI / ASN / LIR pricing model of today and where these costs goes, if "there is not much in actual additional costs" of giving out and maintaining PI resources at discounted prices to some.
What I meant is that maintaining the data once it's in the relevant DBs is effectively free. Yes, there will be costs incurred by extra storage, load on the whois servers etc, but this is all negligible, imo, when compared to the having actual humans interact with other humans.
The real truth is that any subsidizing will be just that; subsidizing.
Of course.
I am in favor of the current simple PI / ASN cost model as it is and oppose subsidizing as it will lead to unbalance. Thus, I want any change in pricing to be discussed for the general PI resource, following today's model.
I am not saying I want anything changed. I simply stated that I think it might be a worthwhile thing to do. I.e. sharing of opinion, not trying to push in any way. Richard
You wrote:
What I meant is that maintaining the data once it's in the relevant DBs is effectively free. Yes, there will be costs incurred by extra storage, load on the whois servers etc, but this is all negligible, imo, when compared to the having actual humans interact with other humans.
I disagree. However well designed, the database system cannot run without some human interaction. People are required to provide technical support, implement new features and do all the regular maintenance you would expect for any operational system. That does not come for free. Regards, Leo Vegoda
On Tue, Apr 19, 2011 at 15:53, Leo Vegoda <leo.vegoda@icann.org> wrote:
I disagree. However well designed, the database system cannot run without some human interaction. People are required to provide technical support, implement new features and do all the regular maintenance you would expect for any operational system. That does not come for free.
It is almost free as a computer will not care in the least if there are a few hundred more or less entries to churn through. The cost comes from maintaining the whole system in the first place. That being said, we are venturing more and more into the mythical lands of off topic so while people are more than welcome to email me privately, I think we should stop this particular sub-thread here. Richard
Hallo Denesh, du schrobst:
On 15 April 2011 13:22, Immo 'FaUl' Wehrenberg <immo.ripe@be.free.de> wrote:
I am working for such a non profit organization voluntarely (as all of our staff does) and I can assure you that fundraising money for servers/routers is not easy for us. When it comes to periodic costs, it is always a trade off between more bandwith/more servers (= increased power costs) and other things (like address space, lir membership fees, etc). So while most end users of independet ressources might not even notice a shift from 50 euros per anno to 500 euros per anno, at least for us that would make things seriously more difficult. After all, thats why we decided not to become a LIR (eventhough it would otherwise fit better then usage of independend ressources). I take your point.. but I have to agree with Jim on this one. I have worked for various non-profit organisations and also understand the financial pressures one is under. However non-profit or not-for-profit does not mean, for all intents and purposes, that it should spend aim to spend more than it can afford.
Indeed not. From my perspective, it should spend exactly as much money as it can afford and to the benefit of the goals it stand for.
Regardless of whether an organisation is for-profit or non-profit, it is a business and must be run like one. Just look at the many charities out there that also run like businesses... and non-profit does not mean it can not make a profit.. just that it has to work on profit-loss neutrality.. and ensure it's costs are met.
My biggest bugbear in some non-profits I have worked in was that the 'real' costs were swept under the carpet... Taking it to it's conclusion, if every volunteer in the organisation was to be paid and all costs were considered, what would be the real cost of running the organisation?
Thats simple. To much. That holds for big non-profit organizations like the Red Cross as well as for small one like the one I work for. I think the point in running such an organization is that it is not to be run like an business, just trying to make as much money as possible, but instead trying to generate as most benifit to the community as possible with the given, limited ressources. Obviously, in that case, reducing the real cost is as important as increasing donations.
Supporting (under-funded) public benefit organisations is all well and
good. But is it fair to expect the NCC membership to pick up the bill? What bill exactly? I don't see much costs generated by users of independent ressources. You mention you are a volunteer in a non-profit organisation. I honestly salute you. However, if you were to charge for your time, how much would the organisation have to pay you?
More than it could afford. That also holds for our other volunteers.
In the same way, whatever resource is being used via RIPE, is a cost to the NCC and that needs to be paid somehow.
I still don't see that. It is a cost to run the NCC, surely. And indeed this cost increases with the number of ressources in use. Without being into finances of Ripe NCC i'd assume that the anual 50 euros per ressource should cover the costs generated by the asignment. CMIIW! Still, PI crowd is supporting their LIR to the usually contracted maintainence fee added to the ressources. regards, Immo Ps: i think we get more and more of topic here, is that ok for this list or should we stop here?
On Fri, 15 Apr 2011, Immo 'FaUl' Wehrenberg wrote:
I still don't see that. It is a cost to run the NCC, surely. And indeed this cost increases with the number of ressources in use. Without being into finances of Ripe NCC i'd assume that the anual 50 euros per ressource should cover the costs generated by the asignment. CMIIW!
For me I don't care where the money goes. A DFZ routing slot is costing money, it's just not costing you money, it's costing "everybody" money, and the cost is that we need routers with bigger FIB sizes the more PIs there are. This has nothing to do with the cost of running RIPE NCC or any other RIR. For me, I don't care if you donate the money to cancer research or physically burn it and spread the ashes, as long as it's costing an adequate amount of money to use this common resource (DFZ slot). -- Mikael Abrahamsson email: swmike@swm.pp.se
Hallo Mikael, du schrobst:
On Fri, 15 Apr 2011, Immo 'FaUl' Wehrenberg wrote:
I still don't see that. It is a cost to run the NCC, surely. And indeed this cost increases with the number of ressources in use. Without being into finances of Ripe NCC i'd assume that the anual 50 euros per ressource should cover the costs generated by the asignment. CMIIW! For me I don't care where the money goes. A DFZ routing slot is costing money, it's just not costing you money, it's costing "everybody" money, and the cost is that we need routers with bigger FIB sizes the more PIs there are. This has nothing to do with the cost of running RIPE NCC or any other RIR.
I am aware of that. Thats the reason oposed to the proposal originally introduced IMHO. However, how would increasing the cost in that case help? Would your saved ripe membership fees compensate that costs? Or would that minority of non-profit users that cannot afford higher PI prices and thus are not able to operate further make any difference? I think for most PI end-users, the cost of pi space is negligable even if it is as high or higher then xs lir fees compared to the costs generated by the provider (including setup and maintainence of the customer network part). I presume, most user would not even notice that costs, thus it would not lead to a significantly less amount of routes in the DFZ. On the other hand, it would make life more difficult for that few non-profit organizations. After all, i don't really see your point here. Immo --
I would say the counter argument to this is that the cost of routing reduces over time, obviously not in line with say Moore's law but to look at the prices of a router that can handle a full ipv4 table compared to 5 years ago, the cost difference is drastically lower. Kind Regards, Find Vaioni's latest Terms and Conditions & SLA docs http://www.vaioni.com/tcaupsla.html. http://www.vaioni.com/http://www.cbawards.co.uk/Reader_Vote_3.cfm http://www.vaioni.com/ http://www.twitter.com/vaioni http://www.twitter.com/vaioni Matthew Hattersley Email: mailto:matthew.hattersley@vaioni.com Mobile: Address: 14 Leslie Hough Way Manchester Lancashire M6 6AJ Tel: 0870-160-0650 Ext 201 Fax: 0870-160-0651 Web: http://www.vaioni.com http://www.convergencesummit.co.uk/Exhibitor_List_North.cfm#Exhibitor_285The information transmitted in and with this email is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Statements and opinions expressed in this e-mail may not represent those of the Company. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender immediately and delete the material from any computer. Please also note, Vaioni filter incoming email for spam and inappropriate words. Unfortunately this does mean that sometimes genuine messages can be filtered out. Although we take measures to recover such messages, it must not be assumed that an email has been received by us and important communications should always be followed up by a phone call, fax or printed copy. -----Original Message----- From: address-policy-wg-admin@ripe.net [mailto:address-policy-wg-admin@ripe.net] On Behalf Of Immo 'FaUl' Wehrenberg Sent: 15 April 2011 15:22 To: address-policy-wg@ripe.net Subject: Re: [address-policy-wg] the cost of PI space Hallo Mikael, du schrobst:
On Fri, 15 Apr 2011, Immo 'FaUl' Wehrenberg wrote:
I still don't see that. It is a cost to run the NCC, surely. And indeed this cost increases with the number of ressources in use. Without being into finances of Ripe NCC i'd assume that the anual 50 euros per ressource should cover the costs generated by the asignment. CMIIW! For me I don't care where the money goes. A DFZ routing slot is costing money, it's just not costing you money, it's costing "everybody" money, and the cost is that we need routers with bigger FIB sizes the more PIs there are. This has nothing to do with the cost of running RIPE NCC or any other RIR.
I am aware of that. Thats the reason oposed to the proposal originally introduced IMHO. However, how would increasing the cost in that case help? Would your saved ripe membership fees compensate that costs? Or would that minority of non-profit users that cannot afford higher PI prices and thus are not able to operate further make any difference? I think for most PI end-users, the cost of pi space is negligable even if it is as high or higher then xs lir fees compared to the costs generated by the provider (including setup and maintainence of the customer network part). I presume, most user would not even notice that costs, thus it would not lead to a significantly less amount of routes in the DFZ. On the other hand, it would make life more difficult for that few non-profit organizations. After all, i don't really see your point here. Immo --
Hi, On Fri, Apr 15, 2011 at 03:21:40PM +0100, Matthew Hattersley wrote:
I would say the counter argument to this is that the cost of routing reduces over time, obviously not in line with say Moore's law but to look at the prices of a router that can handle a full ipv4 table compared to 5 years ago, the cost difference is drastically lower.
Is that so? We paid significantly more for the box that handles 1 million FIB entries as opposed to the box that handles 256k entries. Gert Doering -- NetMaster -- did you enable 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 (89) 32356-444 USt-IdNr.: DE813185279
To get specific, we had a Cisco 7206VXR NPE400 to handle 256k which we bought for around £13k around 6-7 years ago. Last month we bought a Cisco 7606 with a RSP720 to handle the 1m+ mark. Cost us £10k. That's my cards on the table. Kind Regards, Find Vaioni's latest Terms and Conditions & SLA docs http://www.vaioni.com/tcaupsla.html. http://www.vaioni.com/http://www.cbawards.co.uk/Reader_Vote_3.cfm http://www.vaioni.com/ http://www.twitter.com/vaioni http://www.twitter.com/vaioni Matthew Hattersley Email: mailto:matthew.hattersley@vaioni.com Mobile: Address: 14 Leslie Hough Way Manchester Lancashire M6 6AJ Tel: 0870-160-0650 Ext 201 Fax: 0870-160-0651 Web: http://www.vaioni.com http://www.convergencesummit.co.uk/Exhibitor_List_North.cfm#Exhibitor_285The information transmitted in and with this email is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Statements and opinions expressed in this e-mail may not represent those of the Company. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender immediately and delete the material from any computer. Please also note, Vaioni filter incoming email for spam and inappropriate words. Unfortunately this does mean that sometimes genuine messages can be filtered out. Although we take measures to recover such messages, it must not be assumed that an email has been received by us and important communications should always be followed up by a phone call, fax or printed copy. -----Original Message----- From: Gert Doering [mailto:gert@space.net] Sent: 15 April 2011 15:34 To: Matthew Hattersley Cc: address-policy-wg@ripe.net Subject: Re: [address-policy-wg] the cost of PI space Hi, On Fri, Apr 15, 2011 at 03:21:40PM +0100, Matthew Hattersley wrote:
I would say the counter argument to this is that the cost of routing reduces over time, obviously not in line with say Moore's law but to look at the prices of a router that can handle a full ipv4 table compared to 5 years ago, the cost difference is drastically lower.
Is that so? We paid significantly more for the box that handles 1 million FIB entries as opposed to the box that handles 256k entries. Gert Doering -- NetMaster -- did you enable 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 (89) 32356-444 USt-IdNr.: DE813185279
Hi Immo,
I am working for such a non profit organization voluntarely (as all of our staff does) and I can assure you that fundraising money for servers/routers is not easy for us. When it comes to periodic costs, it is always a trade off between more bandwith/more servers (= increased power costs) and other things (like address space, lir membership fees, etc). So while most end users of independet ressources might not even notice a shift from 50 euros per anno to 500 euros per anno, at least for us that would make things seriously more difficult. After all, thats why we decided not to become a LIR (eventhough it would otherwise fit better then usage of independend ressources).
Looking at the cost for PI purely from the viewpoint of a foundation isn't realistic imho. There are plenty of LIR's that actually sponsor the cost of PI to a foundation, especially if it is strategic for the foundation their operations. Not knowing what kind of foundation you are working for, but foundations that come to my mind that would require PI and multihoming: Wikimedia or AMS-IX or alike. Looking at reasonable cost for PI IPv6 (when the multihoming requirement would be removed) in my opinion, should be around 250 euro yearly. Rationale for that 250 euro cost is: . The cost will most likely be enough to remove IT Pet project requests behind a DSL line @home . That kind of yearly cost for PI IPv6 still provides small business and even non-profits the option to start with IPv6 deployment if it is strategic for them to have their own IP's, without hindering IPv6 adoption and deployment. In fact, when the cost for PI IPv6 would increase but the requirement for multihoming is dropped, SMB's & foundations are in a better place than currently as it will be cheaper for them to use PI IPv6 space. They don't have to setup a complete multi-homed environment, but could just ask their 'ISP of the month' to route their IP space for them and use their redundant setup. Running a multi-homing environment isn't for the fainthearted . . . and certainly not cheap or easy. And if you can't make the (strategic/financial) business case to shell out 250 euro yearly for PI IPv6, you can always ask your ISP for regular PA IPv6. Regards, Erik Bais
Hallo Erik, du schrobst:
Hi Immo,
I am working for such a non profit organization voluntarely (as all of our staff does) and I can assure you that fundraising money for servers/routers is not easy for us. When it comes to periodic costs, it is always a trade off between more bandwith/more servers (= increased power costs) and other things (like address space, lir membership fees, etc). So while most end users of independet ressources might not even notice a shift from 50 euros per anno to 500 euros per anno, at least for us that would make things seriously more difficult. After all, thats why we decided not to become a LIR (eventhough it would otherwise fit better then usage of independend ressources). Looking at the cost for PI purely from the viewpoint of a foundation isn't realistic imho.
Agreed, and that was not my intention. However, IMO the non profit organization viewpoint should also be considered, and thats why I joined the discussion.
There are plenty of LIR's that actually sponsor the cost of PI to a foundation, especially if it is strategic for the foundation their operations.
Not knowing what kind of foundation you are working for, but foundations that come to my mind that would require PI and multihoming:
Wikimedia or AMS-IX or alike.
In my case, it is a provider of services for various group, among others political activists. In that case, multihoming and independent ressources means also independence from your ISP (as you have multiple of them and he is not connected to your ressources) and thus it is more likely to get around short handed reactions on law-enforcement-inquires and simmelar political pressure. Thats why we are multihomed and go for our own ressources instead of housing our stuff in a large datacenter with resources from some cheap upstream provider.
Looking at reasonable cost for PI IPv6 (when the multihoming requirement would be removed) in my opinion, should be around 250 euro yearly.
Rationale for that 250 euro cost is:
. The cost will most likely be enough to remove IT Pet project requests behind a DSL line @home
Are there a reasonable number of that? I personally don't see that they are commen, not even in the network geek community.
. That kind of yearly cost for PI IPv6 still provides small business and even non-profits the option to start with IPv6 deployment if it is strategic for them to have their own IP's, without hindering IPv6 adoption and deployment.
From our particular point of view, if I assume 250eur not only for IPv6 PI but also for IPv4 PI and AS numbers, we would be already above the XS membership fee.
And if you can't make the (strategic/financial) business case to shell out 250 euro yearly for PI IPv6, you can always ask your ISP for regular PA IPv6.
It's certainly mostly better then going for PI if you aren't in a multihomed environment. I actually see more points against dropping the multihomed requirement then pro increasing the fees in your reasining. Regards, Immo
On Fri, Apr 15, 2011 at 04:55:45PM +0200, Immo 'FaUl' Wehrenberg wrote:
From our particular point of view, if I assume 250eur not only for IPv6 PI but also for IPv4 PI and AS numbers, we would be already above the XS membership fee.
As far as I remember, ASN is independent resource and is not included in the LIR fee (is extra paid). Still, XS membership fee is 1300 EUR/year which is little bit above 5 times 250 EUR. Piotr -- gucio -> Piotr Strzyżewski E-mail: Piotr.Strzyzewski@polsl.pl
* Piotr Strzyzewski:
On Fri, Apr 15, 2011 at 04:55:45PM +0200, Immo 'FaUl' Wehrenberg wrote:
From our particular point of view, if I assume 250eur not only for IPv6 PI but also for IPv4 PI and AS numbers, we would be already above the XS membership fee.
As far as I remember, ASN is independent resource and is not included in the LIR fee (is extra paid). Still, XS membership fee is 1300 EUR/year which is little bit above 5 times 250 EUR.
And the XS membership fee can possibly apply only to those who have received their IPv4 PA allocation between 2004 and 2008 (and prior to the mid-90s or so). The initial IPv4 allocation is not included in the membership fee. -- Florian Weimer <fweimer@bfk.de> BFK edv-consulting GmbH http://www.bfk.de/ Kriegsstraße 100 tel: +49-721-96201-1 D-76133 Karlsruhe fax: +49-721-96201-99
On Apr 15, 2011, at 4:44 AM, Jim Reid wrote:
On 15 Apr 2011, at 12:04, Michiel Klaver wrote:
Not everything on the internet is a bussiness with needs and a large bag of cash. Please also consider the non-profit organisations with valid reasons of running their own PI network. Supporting (under-funded) public benefit organisations is all well and good. But is it fair to expect the NCC membership to pick up the bill?
Long ago in a different universe, another RIR had a policy that organizations could petition the RIR membership (though the executive council) for fees to be waived. This was seen as a way in which folks (particularly in developing countries) who were unable to afford the registry fees could have those fees subsidized by the membership in an open and transparent manner. Don't know if that policy was actually used... Regards, -drc
On Fri, Apr 15, 2011 at 12:01:46PM +0200, Mikael Abrahamsson wrote:
If someone is willing to spend a substantial amount of money per year to hold a PI, they have a real business need and are willing to back this with cash because it's still cheaper for them. If they don't, then please stop using DFZ resources for no real need.
So "real need" is defined as "commercial interest, backed by spending money for a database entry"? Regards, Daniel -- CLUE-RIPE -- Jabber: dr@cluenet.de -- dr@IRCnet -- PGP: 0xA85C8AA0
On 15 April 2011 10:22, Emilio Madaio <emadaio@ripe.net> wrote:
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
You can find the full proposal at:
Why should a company require PIv6 addressing when the task of renumbering within IPv6 space has become so simple? I'd rather that the size of blocks provided to end users were standardised to reduce the number of possible changes that might be needed, but otherwise this seems to be a request for a change that has no benefit to the operation of the network (even though I can see valid commercial reasons for wanting to do it) We made this mistake with PIv4 please can we not do it again with v6 All that means I'm not in favour of this change. J -- James Blessing 07989 039 476
On Fri, Apr 15, 2011 at 12:28, boggits <boggits@gmail.com> wrote:
Why should a company require PIv6 addressing when the task of renumbering within IPv6 space has become so simple?
Why should a company need to renumber using a simple process when there are easier paths? While this policy change will most likely increase the number of global routes, I personally don't think this should become too much of a problem provided the prices keep the same or even increase. As Michiel Klaver pointed out, special, lower, pricing for NFPs might make sense, but this is way outside the scope of this proposal. Long story short, I am in favour of this change. Richard
On 15 Apr 2011, at 11:28, boggits wrote:
On 15 April 2011 10:22, Emilio Madaio <emadaio@ripe.net> wrote:
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
You can find the full proposal at:
Why should a company require PIv6 addressing when the task of renumbering within IPv6 space has become so simple?
I'm interested: how simple do you think it has become? Tim
Tim Chown wrote:
On 15 Apr 2011, at 11:28, boggits wrote:
On 15 April 2011 10:22, Emilio Madaio <emadaio@ripe.net> wrote:
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
You can find the full proposal at:
Why should a company require PIv6 addressing when the task of renumbering within IPv6 space has become so simple?
I'm interested: how simple do you think it has become?
In particular, when most of the (originally conceived and RFC'd) support functionality in DNS - has been removed in the meantime... Wilfried.
Tim
On Fri, Apr 15, 2011 at 5:22 AM, Emilio Madaio <emadaio@ripe.net> wrote:
I agree with this proposal. Regards, Martin
I support this proposal. Jasper -----Original Message----- From: address-policy-wg-admin@ripe.net [mailto:address-policy-wg-admin@ripe.net] On Behalf Of Emilio Madaio Sent: Friday, April 15, 2011 11:23 AM To: policy-announce@ripe.net Cc: address-policy-wg@ripe.net Subject: [address-policy-wg] 2011-02 New Policy Proposal (Removal of multihomed requirement for IPv6) Dear Colleagues, A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion. You can find the full proposal at: http://www.ripe.net/ripe/policies/proposals/2011-02 We encourage you to review this proposal and send your comments to <address-policy-wg@ripe.net> before 13 May 2011. Regards Emilio Madaio Policy Development Officer RIPE NCC Op dit e-mailbericht is een disclaimer van toepassing, welke te vinden is op http://www.espritxb.nl/disclaimer
On Fri, Apr 15, 2011 at 11:22:35AM +0200, Emilio Madaio wrote:
Although this proposal questionably focusses only on businesses in the rationale, I fully support this proposal. Best regards, Daniel -- CLUE-RIPE -- Jabber: dr@cluenet.de -- dr@IRCnet -- PGP: 0xA85C8AA0
On Apr 15, 2011, at 11:22 AM, Emilio Madaio wrote:
Dear Colleagues,
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
Hello, I support this proposal. Regards, -- Teun Vink, Network & UNIX Engineer BIT BV | teun@bit.nl | +31 318 648 688 KvK: 09090351 | GPG: 0x5A04F4E2 | RIPE: TEUN-RIPE
On Fri, 15 Apr 2011, Emilio Madaio wrote:
We encourage you to review this proposal and send your comments to <address-policy-wg@ripe.net> before 13 May 2011.
As has been stated by me before, 50 EUR PI with no technical requirements for multihoming or other is a recipe for longterm disaster in my book. I strongly oppose. -- Mikael Abrahamsson email: swmike@swm.pp.se
I agree with this proposal. -- Best regards Vegar Løvås Rent a Rack AS On 15.04.2011 11:22, Emilio Madaio wrote:
Dear Colleagues,
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
You can find the full proposal at:
http://www.ripe.net/ripe/policies/proposals/2011-02
We encourage you to review this proposal and send your comments to <address-policy-wg@ripe.net> before 13 May 2011.
Regards
Emilio Madaio Policy Development Officer RIPE NCC
Dear address-policy-wg, I would like to express my support for this policy proposal. The promise of auto-configuration making transitioning between v6 prefixes seamless is not yet fully delivered, and eliminating the manual renumbering overhead when moving between service providers, as well as having a clear path to multihoming ("just add ASN") should assure a lively Internet services market for years to come. -- Respectfully yours, David Monosov On 04/15/2011 11:22 AM, Emilio Madaio wrote:
Dear Colleagues,
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
You can find the full proposal at:
Dear All, I agree with removing the multi-homing requirement for IPv6 PI. Its pretty awkward to send your customers to a competitor because to deploy IPv6 PI space he or she needs to be multi-homed. Also, rising technologies such as LISP allow end-users to be multi-homed in a way that is transparent to the DFZ, so why bother restricting people to BGP multi-homing. Kind regards, Job Snijders On 04/15/2011 11:22 AM, Emilio Madaio wrote:
Dear Colleagues,
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
You can find the full proposal at:
Considering the arguments from both David and Job, I would also like to express my support. With kind regards, Michiel Klaver At 06-05-2011 12:40, Job Snijders wrote:
Dear All,
I agree with removing the multi-homing requirement for IPv6 PI.
Its pretty awkward to send your customers to a competitor because to deploy IPv6 PI space he or she needs to be multi-homed. Also, rising technologies such as LISP allow end-users to be multi-homed in a way that is transparent to the DFZ, so why bother restricting people to BGP multi-homing.
Kind regards,
Job Snijders
At 06-05-2011 12:28, David Monosov wrote:
Dear address-policy-wg,
I would like to express my support for this policy proposal.
The promise of auto-configuration making transitioning between v6 prefixes seamless is not yet fully delivered, and eliminating the manual renumbering overhead when moving between service providers, as well as having a clear path to multihoming ("just add ASN") should assure a lively Internet services market for years to come.
-- Respectfully yours,
David Monosov
I'm supporting this proposal, as policy for IPv6 PI should be similar to existing IPv4 PI policy and similar rules should be applied for request processing. With regards, Daniel On 04/15/2011 11:22 AM, Emilio Madaio wrote:
Dear Colleagues,
A proposed change to the RIPE Document ripe-512,"IPv6 Address Allocation and Assignment Policy", is now available for discussion.
You can find the full proposal at:
http://www.ripe.net/ripe/policies/proposals/2011-02
We encourage you to review this proposal and send your comments to <address-policy-wg@ripe.net> before 13 May 2011.
Regards
Emilio Madaio Policy Development Officer RIPE NCC
participants (27)
-
boggits -
Daniel Roesen -
Daniel Suchy -
David Conrad -
David Monosov -
Denesh Bhabuta -
Emilio Madaio -
Erik Bais -
Florian Weimer -
Gert Doering -
Hank Nussbacher -
Immo 'FaUl' Wehrenberg -
Jasper Jans -
Jim Reid -
Job Snijders -
Leo Vegoda -
Martin Millnert -
Matthew Hattersley -
Michiel Klaver -
Mikael Abrahamsson -
Piotr Strzyzewski -
Richard Hartmann -
Sander Steffann -
Teun Vink -
Tim Chown -
Vegar Løvås -
Wilfried Woeber, UniVie/ACOnet