2007-08 review phase comments
Dear colleagues ETNO's recently submitted a contribution, titled "IP Addressing in a post IPv4 World-Principles", to the RIPE address-policy-wg mailing list (available also to the ETNO web site, http://www.etno.be/Default.aspx?tabid=2058). This contribution provides an analysis on why, according to ETNO Members, any emerging marketplace should not determine IPv4 address exhaustion management. Quoting some basic arguments from the abovementioned contribution, ETNO believes that it is the successful bottom-up processes that keep the Internet community from being engaged in discussions regarding intervention or new models of governmental control and therefore it is the Internet community that should take all the appropriate measures in order to adhere to the current governance model. Facilitating a market based on the principle that attaches an intrinsic "value" to an IP address would engage competition authorities, policymakers and will raise legal issues. This should be carefully considered before introducing policies that will facilitate a market. Therefore, ETNO has some strong concerns regarding the possible impacts of the introduction of a transfer option or an open market option. This does not mean that ETNO fails to recognise the possibility that future transactions that will involve transfers of IP addresses between entities for profit. It means that internet community will need to carefully consider the implications of assisting such a process, taking as a starting point that the Internet community wants to preserve the current internet governance bottom-up processes. If a market emerges, future developments relating to IPv4 addresses must not undermine the viability of the Internet. Therefore based on the above, ETNO members could not support the 2007-08 Policy Proposal, Enabling Methods for Reallocation of IPv4 Resources. Christina Kelaidi ETNO Naming Addressing and Numbering Issues WG.
Quoting some basic arguments from the abovementioned contribution, ETNO believes that it is the successful bottom-up processes that keep the Internet community from being engaged in discussions regarding intervention or new models of governmental control and therefore it is the Internet community that should take all the appropriate measures in order to adhere to the current governance model. Facilitating a market based on the principle that attaches an intrinsic "value" to an IP address would engage competition authorities, policymakers and will raise legal issues. This should be carefully considered before introducing policies that will facilitate a market. Therefore, ETNO has some strong concerns regarding the possible impacts of the introduction of a transfer option or an open market option. This does not mean that ETNO fails to recognise the possibility that future transactions that will involve transfers of IP addresses between entities for profit. It means that internet community will need to carefully consider the implications of assisting such a process, taking as a starting point that the Internet community wants to preserve the current internet governance bottom-up processes. If a market emerges, future developments relating to IPv4 addresses must not undermine the viability of the Internet.
Christina, You're not proposing much of an alternative here, other than hoping that things are just going to sort-of carry on the way they always have. It would be nice to believe that this was going to happen, but personally, I don't view this as a realistic potential future. ETNO is correct in identifying that a market based addressing approach will attract lots of attention from forces which are - to be frank - unwanted and largely unhelpful. However, it must also be realised that an IP address market already exists where you can buy and sell IPv4 address space. And in a world where there is no longer a free supply of nearly zero-cost address space, this market will take off at high speed, regardless of whether or not the RIRs choose to engage in it. If they choose not to, I would argue that the effects on the marketplace will be significantly worse than if they choose to join and continue to play their part as registrars of data. Chaos is the most likely outcome if they are not given a mandate to be involved; RIRs will not have the community mandate to ensure that their databases of address holders continue to correspond with the ultimate users of the address space. This will lead to significant problems in many areas, not least technical and legal. Unless the RIRs are allowed to adapt to the de-facto methods of managing address space, they will become an irrelevant part of ipv4 address management, and we will see significantly more regulatory, political and legal interest being expressed than if they are given a mandate to facilitate address transfer. I don't think that there are very many people who really want an address space market; however, given the circumstances, it would seem to me and many others that it is the least bad of a small number of problematic ways of dealing with a severe constriction of IPv4 address space supply.
Therefore based on the above, ETNO members could not support the 2007-08 Policy Proposal, Enabling Methods for Reallocation of IPv4 Resources.
Christina, your email above has stated ETNO's position but nothing more. However, it is clear that ETNO has debated this issue in some depth. It would be helpful to this process if we could understand your reasoning more. In the interests of helping people in the RIPE community to understand the ETNO position, would it be possible for you to give details about why you have taken this position? Nick -- Network Ability Ltd. | Head of Operations | Tel: +353 1 6169698 3 Westland Square | INEX - Internet Neutral | Fax: +353 1 6041981 Dublin 2, Ireland | Exchange Association | Email: nick@inex.ie
You're not proposing much of an alternative here, other than hoping that things are just going to sort-of carry on the way they always have. It would be nice to believe that this was going to happen, but personally, I don't view this as a realistic potential future.
I don't think anybody believes that things are always going to be the same. The point that I think ETNO is trying to stress is that RIPE should be careful to not extend its scope into areas that are normally the responsibility of government regulators and this includes market regulators. I think that we all want RIPE to continue the process of industry self regulation in IP addressing because the IPv4 shortage is just a temporary problem that will go away when IPv6 is widely deployed. Rather than focus on a shortage problem which RIPE can *NOT* solve, we should focus on other issues where RIPE is a valuable forum.
ETNO is correct in identifying that a market based addressing approach will attract lots of attention from forces which are - to be frank - unwanted and largely unhelpful. However, it must also be realised that an IP address market already exists where you can buy and sell IPv4 address space. And in a world where there is no longer a free supply of nearly zero-cost address space, this market will take off at high speed, regardless of whether or not the RIRs choose to engage in it.
This proposal enables the sale of IP addresses as property, even though it keeps the financial details secret. This legitimizes the position that IP addresses, and phone numbers are assests which can be bought and sold. But the policy as written places restrictions on the market which cannot be justified. In the first place, if IP addresses are an asset, then RIPE has no authority to say who can buy or sell these assets. Such restrictions can only be made by governments, and RIPE does not have that authority. I would rather see RIPE reject such proposals because it reinforces the message that IP addresses are not property which can be bought and sold.
Chaos is the most likely outcome if they are not given a mandate to be involved; RIRs will not have the community mandate to ensure that their databases of address holders continue to correspond with the ultimate users of the address space.
The only way that this chaotic situation could come to pass would be if many ISPs of a significant size, started to buy and sell IP addresses. ETNO contains many of the largest ISPs in Europe and they say that they will not do this. In addition, we are entering a period of IPv4 address shortage which means that most ISPs of any size will not have any IP addresses to sell unless they can get a price significantly above what they can earn by using the IP addresses to sell services. These are very real forces and they serve to limit chaos. I would prefer for RIPE to wait and see if there is increasing evidence of descent into chaos before taking actions like 2007-08.
Unless the RIRs are allowed to adapt to the de-facto methods of managing address space, they will become an irrelevant part of ipv4 address management,
Adaptation is good, just not in the directions defined by 2007-08. For instance, 2007-08 allows the addresses to be transferred without disclosure of any of the details of the transaction, for instance price paid. This lack of transparency would lead to addressing cartels and is not consistent with the principles of OPEN self regulation.
I don't think that there are very many people who really want an address space market; however, given the circumstances, it would seem to me and many others that it is the least bad of a small number of problematic ways of dealing with a severe constriction of IPv4 address space supply.
Seems to me that the least bad way is to deploy IPv6. There are at least two or three years left to do this, and it does not require replacing very much infrastructure. No need to panic. --Michael Dillon
michael.dillon@bt.com wrote:
this includes market regulators. I think that we all want RIPE to continue the process of industry self regulation in IP addressing because the IPv4 shortage is just a temporary problem that will go away when IPv6 is widely deployed.
With respect, there is no guarantee that ipv6 will be taken up as a solution to the ipv4 depletion problem. It is certainly one option, but there are others such as provider NAT and LISP. Who knows, there may be multiple solutions deployed? In fact, it's quite likely that there will be many different approaches taken.
This proposal enables the sale of IP addresses as property, even though it keeps the financial details secret.
This will happen regardless of whether the RIRs get involved or not.
The only way that this chaotic situation could come to pass would be if many ISPs of a significant size, started to buy and sell IP addresses. ETNO contains many of the largest ISPs in Europe and they say that they will not do this.
Michael, I really don't believe for a moment that ETNO members are going to keep away from the IPv4 market when it comes into being, regardless of the current ETNO position in these times of plenty. While you and other people like you and me are - to one degree or another - against the whole notion of a market, if a shortage of IPv4 address space starts taking its toll on a company's financials, the decision to purchase more address space will be taken away from the technical departments and assumed by the business development departments (and perhaps your boards of management/directors) in your companies.
In addition, we are entering a period of IPv4 address shortage which means that most ISPs of any size will not have any IP addresses to sell unless they can get a price significantly above what they can earn by using the IP addresses to sell services. These are very real forces and they serve to limit chaos. I would prefer for RIPE to wait and see if there is increasing evidence of descent into chaos before taking actions like 2007-08.
By the time chaos arrives, it will be too late. Rescuing oneself retroactively from chaos is much more difficult than taking a more proactive approach from the beginning.
Adaptation is good, just not in the directions defined by 2007-08. For instance, 2007-08 allows the addresses to be transferred without disclosure of any of the details of the transaction, for instance price paid. This lack of transparency would lead to addressing cartels and is not consistent with the principles of OPEN self regulation.
I would see it as likely that most ipv4 address space will be auctioned using either a highest bid (i.e. most fully open auctions) or a perturbed second highest bid approach (i.e. the ebay auction method) And again, whether or not the prices are disclosed is completely orthogonal to whether the RIRs get involved.
I don't think that there are very many people who really want an address space market; however, given the circumstances, it would seem to me and many others that it is the least bad of a small number of problematic ways of dealing with a severe constriction of IPv4 address space supply.
Seems to me that the least bad way is to deploy IPv6. There are at least two or three years left to do this, and it does not require replacing very much infrastructure. No need to panic.
This is off-topic for APWG, but IPv6 deployment will almost certainly not happen until the latest possible moment. And there are other more important and more costly issues associated with deploying an ipv6 internet than just rolling out new hardware and coming up with an addressing plan. Nick -- Network Ability Ltd. | Head of Operations | Tel: +353 1 6169698 3 Westland Square | INEX - Internet Neutral | Fax: +353 1 6041981 Dublin 2, Ireland | Exchange Association | Email: nick@inex.ie
Nick, Even if there will be a market to buy Ipv4 address space, it will not be able to satisfy the address needs of the bigger broadband providers. (ETNO or other) Hoping that a IPv4 maket will give you the required addresses to continue to assign an Ipv4 address to each (new) individual subscriber is wishfull thinking. So we (as an industry) have to find a cost effective solution to consume (much) less than 1 IPv4 address per broadband user between now and exhaustion time. (Ipv6 or NAT or other) Migrating existing customers to a new solution/service is not easy or cheap (especially if they have to replace their CPE). I suppose that this solution (IPv6) will be apllied first to the growth (new users) and later to migrate existing users. Marc
The only way that this chaotic situation could come to pass would be if many ISPs of a significant size, started to buy and sell IP addresses. ETNO contains many of the largest ISPs in Europe and they say that they will not do this.
Michael, I really don't believe for a moment that ETNO members are going to keep away from the IPv4 market when it comes into being, regardless of the current ETNO position in these times of plenty. While you and other people like you and me are - to one degree or another - against the whole notion of a market, if a shortage of IPv4 address space starts taking its toll on a company's financials, the decision to purchase more address space will be taken away from the technical departments and assumed by the business development departments (and perhaps your boards of management/directors) in your companies.
__________________________________________________ Marc Neuckens belgacom Manager Expert Center IP & Internet ITN/IRO/Remote Network Operations(RNO)/ECI Lebeaustraat 2 B-1000 Brussel Tel.: +32-2-514.43.66 E-mail: <marc.neuckens@belgacom.be> __________________________________________________ **** DISCLAIMER **** http://www.belgacom.be/maildisclaimer
At 00:04 29/05/2008, marc.neuckens@belgacom.be wrote:
Nick,
Even if there will be a market to buy Ipv4 address space, it will not be able to satisfy the address needs of the bigger broadband providers. (ETNO or other) Hoping that a IPv4 maket will give you the required addresses to continue to assign an Ipv4 address to each (new) individual subscriber is wishfull thinking.
So we (as an industry) have to find a cost effective solution to consume (much) less than 1 IPv4 address per broadband user between now and exhaustion time. (Ipv6 or NAT or other)
Migrating existing customers to a new solution/service is not easy or cheap (especially if they have to replace their CPE). I suppose that this solution (IPv6) will be apllied first to the growth (new users) and later to migrate existing users.
We just finished implementing a procurement for upstream. One thing we found with many of the really large ISPs is their poor support for IPv6. Few if any had IPv6 throughout their backbones. We should all be pushing our peers to provide a complete v6 service. That's the only way it's going to be available when v4 space really becomes tight. When your local provider goes rushing to his upstream, they'd better be able to provide a v6 service. ISPs could start marketing themselves as "v6 ready" rather like was done for the millenium. -- Tim
We just finished implementing a procurement for upstream. One thing we found with many of the really large ISPs is their poor support for IPv6. Few if any had IPv6 throughout their backbones.
Don't confuse poor demand with poor support. No company can afford to implement IPv6 in areas where there is no demand for it. As a result I think you will find that all the large network providers who supply IPv6 services are building out IPv6 as the demand rises. This is especially true of network using MPLS since 6PE only needs to be turned on for the few edge routers with IPv6 customers connected.
We should all be pushing our peers to provide a complete v6 service.
And those peers will ignore you since there is no business case to deploy IPv6 everywhere.
That's the only way it's going to be available when v4 space really becomes tight.
Not true. If a network provider has proven their IPv6 network design in the lab, and has a detailed plan for deployment, they will be able to get IPv6 out there in a matter of months. In addition, it makes sense to deploy the OSS and NMS support for IPv6 before turning it on in the network. This kind of activity is invisible to the outside world, but is essential in being ready for IPv4 exhaustion.
ISPs could start marketing themselves as "v6 ready" rather like was done for the millenium.
I have no doubt that we will see this in about two years, and some companies will suffer severe financial pain because they are not ready. --Michael Dillon
participants (5)
-
Kelaidi Christina -
marc.neuckens＠belgacom.be -
michael.dillon＠bt.com -
Nick Hilliard -
Tim Streater