[rpd] IPv4 Soft Landing BIS

[Owen DeLong]

> On Jul 28, 2017, at 09:58 , Noah <noah at neo.co.tz> wrote:
> 
> 
> 
> On 28 Jul 2017 10:01 a.m., "Owen DeLong" <owen at delong.com <mailto:owen at delong.com>> wrote:
> 
>> On Jul 27, 2017, at 23:17 , Noah <noah at neo.co.tz <mailto:noah at neo.co.tz>> wrote:
>> 
>> Hi Community, 
>> 
>> Lets put our chief opposers of the softlanding policies aside for a second..
> 
> Yes… This is a typical tactic that you have repeatedly attempted. Let’s brush aside anyone who doesn’t agree with us and pretend that their arguments either don’t exist, have no merit, or can otherwise simply be ignored.
> 
> 
> No no Owen. Opposition is good dont get me wrong which is why I clearly and openly crown you and Andrew as opposers while i am a supporter. 

Yes, but I have never suggested putting aside your views. I have argued against them, but I have never suggested that they should be set aside or ignored. 

> Unfortunately for you, the co-chairs don’t have this option. They must consider all community input no matter how much authors don’t like it.
> 
> 
> Co-chairs will determine between opposers and support what is rational and determine for us the way forward. So again dont get me wrong.

Perhaps I have misread the PDP. Please point me to the part where it says that the co-chairs are responsible for determining what is rational.

I thought (apparently mistakenly) that their job was to determine if there is consensus. Since there remain supporters and opposers, there is clearly not consensus. Consensus occurs when the objections raised by the opposers have been fully addressed and either resolved or agreement is reached that the proposal can move forward without addressing the objections.

Since none of the opposers have agreed to moving forward without having their objections addressed, I think it will be very hard, indeed, for the co-chairs to come to the conclusion that there is community consensus for this proposal.

>> I am not sure if I will make sense but let me try to present a case below.
> 
> This is simply too easy and rather than comment, I will let the statement stand on its own merits for others to interpret as they see fit.
> 
>> We supporters of an updated Softlanding policy believe in the impact that a public IP address can have on our peoples lives.
> 
> So do the opposers.
> 
> Ack
> 
> 
>> Its a fact that an IP address has been responsible for the creation of jobs and most of you if not all of you in this community in one way or another are positively affected by the internet.
> 
> Yes… So let’s not deny those abilities to real providers creating real networks and services today in favor of keeping them on the shelf for some future provider that may or may not even actually exist in this imaginary future.
> 
> 
> No one is going to deny any existing LIR or PI member the ability to apply for additional resorces. 

True… The proposed policy would not prevent them from applying. On this we agree. Unfortunately, it would actually prevent them from receiving the resources they need and this is why we oppose the current draft of the policy.

> In the grand scheme of things, everyone would be able to apply for the resources and get them on a need basis.

The first part of this statement is true. The latter part is specifically not true under this proposal.

Specifically, the second part of the statement is contrary to 5.4.3.1, 5.4.3.2, 5.4.5, and most especially 5.4.6.

(source: https://www.afrinic.net/en/community/policy-development/policy-proposals/2153-ipv4-soft-landing-bis for those who wish to review it)

> 
> 
> 
> 
>> Our governements though the LIR's and PI resource members of  Afrinic collect taxes from this businesses hence develop our respective countries.
> 
> You are again continuing to make the same case that we opposers are making against this policy. Perhaps we don’t disagree so much as I thought..
> 
> Like i said, i have my views and you have yours.

Difficult for me to understand how this is a response to my statement.

>> The IP address has enabled egoverment and now the administrative state can serve citizens through egov services thereby increase government effeciency and more and more of our African governments are putting information online.
> 
>> The IP address is responsible for research and education institutes today across the continent to communicate and share ideas and our scholars are able to interact by means of the African Internet.
>> 
>> 
>> The IP address has enabled access to information beyond our reach and this access to information is enabling our people effect socio-economic and political change.
>> 
>> 
>> The IP address has enabled ecommerse and has empowered local fintech innovations like mobile money which has fundamentally had a profound effect on our peoples lives.
>> 
>> The Internet that is a social media enabler has enabled young men and women across this continent to access a local market and trade online on ecommerce platforms where they dont have to own physical stores/shops but they can market and sale their products online and deliver the same to the buyers.
>> 
>> With a surge in high unemployments rates across sub-sahara Africa, many unemployed graduates are finding a reason to hope as platforms like Instagram, facebook and others are enabling them to access followers who in one or another have turned out to be a customer or client base.
>> 
>> Our local musicians today are reaching a far bigger audience that has enabled them grow their artistic talent beyond Africa and our Afrobeat music and house music is now listened and enjoyed by folks from all works of life. They are enabling us change our story.
>> 
>> The IP address bas enables our friends who come to tour our beautiful continent access to more local information about our various countries thereby enabling us collect revenue and create jobs and build our hospitality industry.
> 
> On all of this we agree. These arguments are, in fact, the best drivers for my opposition to this proposed policy because it will interfere with the continued deployment of all of these things.
> 
> How will it interfear yet we are saying resources will continue to be allocated under this policy on need basis as has been but with some mechanism in place that enables others who may need also get the resources.

While you are saying that resources will continue to be allocated, you are making this look black and white while there are many shades of grey available.

This proposal (as stated above) places significant restrictions and reductions on this allocation process and places existing networks at an extreme disadvantage. It prevents existing network operators from getting additional space more than once every 2 years, while only allowing them to get 8 months worth of address space, for example.

There are other limitations as well, some even more severe for some providers.

> When Afrinic opened doors back in 2004, did everyone on this continent get all resources at the same time?
> 
> There are startups that cropped up later in the past 5 years with the arrival of the submarine fiber and have been able to also get some resources because the process was always on need basis and not on rush rush and deplete.

Sure, but I don’t see how that justifies saying no to providers that exist today and creating an artificial shortage today to prevent providers that don’t exist from experiencing that shortage many years in the future. It’s this part of the policy proposal that simply doesn’t make any sense to me. Can you please explain the logic of doing that to me?

>> Lets not kid ourselves. As the vendors continue fixing the IPv6 software stacks, IPv4 still works and IPv4 will not be less useful anytime soon because vendors across the board still make a killing out of IPv4. 
> 
> Um? You do realize that IPv4 doesn’t actually work significantly better than IPv6 in the vast majority of software stacks, right?
> 
> 
> Ack that IPv6 is superior improvement but that is besided the point. The lesser IPv4 is the status quo today and will remain as such in this Africa continent because Africa is not India or China but more than 50 sovereign states each with its own local laws and socio-economic  and political policies.

So you are advocating for Africa to once again become disenfranchised from the world internet by being held back by this disadvantageous protocol even as the rest of the world moves forward with IPv6? Isn’t this the kind of attitude that would get someone of my skin color branded as a colonialist?

You were the one who started with claiming that IPv4 still works while we are fixing IPv6. If you can’t remember that, I have retained your quote above my earlier statement here for context. (3 levels of quotation above my 2-level).


> You do realize that when IPv4 was only 20 years old, it suffered from even more grievous problems than the worst problems with IPv6 today, right?
> You do realize that IPv6 works well enough that Facebook, Google, Comcast, and every major cloud provider are now supporting it, right?
> 
> 
> I do realise all that but let me ask you one quastion.
> 
> Do you think facebook or Google whose business is based on numbers will look aside and ignore a African market that has more than 1billion people that still access their services using IPv4.

When the cost of maintaining an IPv4 infrastructure exceeds the value of those customers, I think they will do it in a heartbeat. However, I don’t believe that more than
1 billion Africans will still be IPv4-only in even 3 years. I believe that many of them will have IPv6, so the question becomes what fraction of Africa gets ignored at that
time more than it is a question of ignoring all of Africa.

Further, that fraction is a monotonically decreasing fraction of the continent, making it progressively less and less valuable to these providers.

> They are earlier adopters and late adopters will also catch up as factors that affect USA based companies are not the same factors that affect Afrinican based companies.
> 
> Africa is a completely different market than America or Europe for that matter.

I keep hearing this, but it doesn’t hold up to scrutiny quite as well as you seem to think.

> Claiming “IPv4 still works” as if “IPv6 doesn’t” is absurd at best and disinformation at worst.
> 
> 
> IPv4 works and I am sending you this email from my IPv4 capable android handset whose software stack unfortunately doesnt have support for IPv6 nor does my date provider.
> 
> In fact even in the office, this phones doesnt support IPv6 of the wifi even thought my macbook pro does support and gets IPv6 addresses through the same dual-stack wifi network.

Wow… Must suck to be you… Sorry to hear that. IPv6 works too and I’m sending you this email from my IPv6 iMAC via my IPv6 enabled Linux Mail server running a very old version of Sendmail on Fedora.

I initially read your email on my IPv6 enabled iPhone which doesn’t even receive an IPv4 address from my cellular provider any more. It is 100% IPv6 only now. Any access to IPv4 sites or content is via NAT64.


> Actually, I don’t think it’s going to be all that long before IPv4 is less useful.
> 
> 
> Yes in North America but not in the African continent with over 50 states with different economic conditions.

So you believe it is good for Africa to again become isolated from the rest of the world? Really?

This is a global internet. If the rest of the internet starts turning off IPv4, then Africa either turns on IPv6 or becomes disconnected from the world. It’s as simple as that.

> You see, big eyeball providers are already starting to see the excessive costs of IPv4 and the cost reductions that are possible with IPv6 monostack. Once a little more critical mass is available on the content side, the cost-benefit equation will actually lead to many of them finding ways to discourage their customers from using IPv4 and/or turning off IPv4 services for those customers.
> 
> 
> Ack and we all wait for that time. Africa is over a billion, China is over a billion and india is over a  billion.

Sure, but if you look at it from an economic perspective, the US eyeball providers don’t need to continue to provide access to China, Africa, or India because the eyeball users on IPv4-only in those countries aren’t able to communicate directly with the eyeballs in the US today anyway. So that’s not a useful economic reason for the US providers to continue to maintain IPv4 connectivity. What’s currently keeping US providers from turning off IPv4 mostly amounts to 4 forces:

	1.	Amazon
	2.	AWS
	3.	Azure
	4.	Google Cloud

At least two of those 4 have (poorly) implemented IPv6 solutions today. The other two are definitely working on it and I expect to see results in less than a year from at least one of them.

> At the very least, as IPv4 moves to more and more CGN and centralization, greater concentration of fate sharing, and a continuing reduction in performance due to centralization and layers of NAT, IPv4 will start to become less and less usable in the very near future. If you read the State of the Internet report you will see that these impacts are already measurable. If you want, there are other reports from Potaroo and other sources that also show this.
> 
> 
> Owen i have read this reports believe me but am telling you that your reality is not my reality.

Interesting… My reality is built from documented proven facts and statistics.

What is your reality built from?

Are you really suggesting that the performance and quality of IPv4 service is not degrading as the deployment and use of NAT increases? Really? Your reality seems most interesting, indeed.

>> Big telecoms have invested in legacy equipmemts and still returning their investments.
> 
> If your equipment is less than 10 years old, you can probably put IPv6 on it. You might need a software upgrade.
> If your equipment is less more than 10 years old but less than 20, you may be stuck with an inappropriate depreciation schedule which makes it more difficult, but that’s a management problem that shouldn’t be all that hard to address if you simply go fess up to the CFO with accurate information. Certainly starting now will turn out better for you than starting later.
> 
> Different enviroment and different economies.

In what way? How does Africa differ from my statement above?

>> Big internet companies that depends on global numbers which are mainly still accessible via the IPv4 internet wont risk loosing this market.
> 
> If the cost exceeds the benefit, you bet they will. IPv4 represents a rising cost per service in an environment of declining revenues.
> 
> 
> Ack
> 
> 
>> Some startups will be seeking addresses to atleast support critical infrastucture like DNS and Web servers if there is none from Afrinic, you be forced to pay more expensively from those who have the space.
> 
> This is a simple reality. Denying addresses to present need in order to protect these johnny-come-lately startups that may not even materialize is absurd. Further, the cost of IPv4 will eventually drive startups to ignore the IPv4 market in favor of simply implementing IPv6 only. There’s no harm in this. There’s greater innovation and better services possible to consumers on IPv6.
> 
> 
> No one is denying anyone addresses. Everone gets on need basis.

This simply isn’t true. Please reread the sections of your proposal referenced above. You will see that they invalidate your claim here.

>> Facebook, youtube,uber,twitter,google,instagram,apple,microsoft,airbnb to name but a few are software companies that have created jobs and are making millions of dollars and affecting lives because the Ip addresses made it possible.
> 
> Nearly every name on that list is operating on IPv6. Facebook is aggressively turning off IPV4 wherever they can.
> 
>> The IP address is what makes you and me pay our bills within our domain of ICT.
> 
> Really? I challenge you to prove this. For me, the transit service I receive from my provider is what makes me pay my ISP bill. I don’t even get my addresses from them.
> [Technically I get one IPv4 address from them which I use to terminate tunnels and a handful of IPv6 addresses which I don’t actually use at all yet. However, I will be migrating my tunnels to v4/v6 over GREv6 eventually]
> 
>> There million dollar IPv4 transfer market is here to stay and late entreprenuers and startups will pay more expensively for an IP address post IPv4 exhaustion which could potentially discourage investment in the IP related investments.
> 
> The million dollar IPv4 transfer market is actually already starting to show significant reductions in availability and in order for a market to continue, you need not only demand (which I believe will somewhat remain), but you also need supply. Now it’s true that supply will ebb and flow with increases in the price per address, but at some point, the cost of an IPv4 address will exceed the ROI and even the last hold-outs will recognize that it’s simply too expensive to continue ignoring IPv6.
> 
>> The last IPv4 address ought to be handled with care hence further recommendation on how to deal with this depletion scenario.
> 
> Again, I agree with the sentiment, but this proposal is the opposite of proper care.
> 
>> We can deplete fast but if any one of you tomorrow wants IP addresses and cant get them from Afrinic, you will be forced to dance to the IPv4 brokers pricing.
> 
> Actually, the brokers aren’t setting the prices. The brokers are middle-men taking a commission on each transaction. It is the people who are willing to free up addresses from their networks and make them available who are setting the price.
> 
> 
> Brokes dont and will never benefit us. Afrinic does benefit the contient and its members.

First, I’m not sure this first part is a true statement. However, I’m not sure how brokers are relevant to this policy proposal. An IPv4 market place is a reality whether policy supports it or not. Brokers are simply a form of matchmaking service within that marketplace. If you find the addresses you need without one, more power to you. If you enlist the help of a broker in the process, that’s fine too. It’s really not relevant to the policy discussion at hand.

I notice you didn’t respond to several points above this one. Does that mean you concede those points?

>> You will be forced to look up to IPv4 brokers for address space and the IPv4 broker will not trade space the same way Afrinic does.
> 
> Actually, if you can find someone who will make addresses available to you by sale or trade without involving a broker, you are free to do so. Many such transactions have occurred in at least 3 of the other RIRs.
> 
> To the benefit of brokers.

How is this to the benefit of brokers?

I’m talking about transactions that occurred between two parties without any broker involved in the process. No broker made any money on these transactions.

> Perhaps you should check your facts before posting such inaccurate data.
> 
> an example fyi 
> 
> https://www.apnic.net/manage-ip/manage-resources/transfer-resources/transfer-facilitators/ <https://www.apnic.net/manage-ip/manage-resources/transfer-resources/transfer-facilitators/>
> 
> https://laruscloudservice.net/offer-ip-for-sell <https://laruscloudservice.net/offer-ip-for-sell>
> 
> http://www.hilcostreambank.com/ipv4-brokerage <http://www.hilcostreambank.com/ipv4-brokerage>

These are examples of brokers. They have nothing to do with the transactions I’m talking about.

Yes, there are brokers. I never said there were not.

Brokers are like a dating service for IP addresses. They help buyers find sellers and vice versa. If you want to use them when you need to buy or sell addresses, you are free to do so, but this is not required.

If you want to buy and you can find a qualified seller without a broker, the two of you can work out your deal and no broker need be involved. Same is true if you want to sell and  can find a qualified buyer.

You are acting as if someone is forcing you to use a broker if you want to make use of the transfer market. THere’s no policy in any RIR that requires such a thing. You are free to pursue your own deals without involving brokers if you wish.

>> IPv4 is still relevant than you will ever imagine save for the rhetoric around how IPv4 is legacy and not needed meanwhile millions of dollars are exchanging hands in the IPv4 tranfer market.
> 
> IPv4 is relevant today. In 2 years, I think its relevance will begin to decline sharply. In 5 years, I think we will see a very different environment. Ignoring this fact and basing policy decisions for the future only on the past and present is a perilous venture.
> 
> 
> So you assume. No one is ignoring anything. We run an extensive network in Africa and I am on the ground in African and i understand what am talking about from the perspective of Africa.

So do others. Akamai has rather extensive deployments in Africa as well. I’m afraid you don’t get to dismiss my claim as mere assumption from an outsider in this case. I realize it would be convenient if you could, but it isn’t.

> But your assumption above is based on your own experience in North America and I am telling you this are 2 different enviroments.

Actually, it’s my projection based on my experience operating in Africa _AND_ Europe _AND_ Asia _AND_ Latin America, Australia, Caribbean, Pacific Islands, …

> America is a country, Africa is more than 50 soverign states.

No, America is roughly 61 countries. The united States is one country in America (made up of 50 sovereign states as it happens).

North America is actually not only the US, Canada, and Mexico, but also many sovereign nations in the Caribbean (as well as a number of dependent territories).

However, America also includes central and south America which add 12 sovereign nations and 4 dependent territories to the count.

While I respect that you have a longer history with operations in Africa, I am _NOT_ as you claim unfamiliar with the environment and my experience spans the globe, not just the US or even North America. I have substantial experience with operations on 6 continents and a substantial history of participation in all 5 RIRs policy processes. I am quite familiar with the policy and operational environments throughout the world.

> 
> 
> 
>> I know its a long one and i could go on but these are my 2cents community.
> 
> So it seems you support and I oppose this proposal for many of the same reasons. This is actually good news. It’s not uncommon for people of good conscience to come to different conclusions presented the same facts based on the difference in their experiences.
> 
> 
> Ack
> 
> 
> It appears that you have some errant beliefs about how the transfer markets work and about the future of IPv4. Perhaps if you take a wider and more objective view of the inherent realities of the economics of IPv4 (suggest you review Lee Howard’s presentation on Per User Per Year costing of IPv4 from the Denver INET meeting:
> 
> https://www.youtube.com/watch?v=vXf8ZIew1j0 <https://www.youtube.com/watch?v=vXf8ZIew1j0>
> 
> This is from 2013 and his projections haven’t been 100% realized yet, but only the rate of cost increase differs. All of the trends are going in the directions he predicted, just somewhat slower than forecast.
> 
> Again, he is speaking from the perspective on the enviroment he is familiar with.

First, I think it isn’t fair to characterize the wide variety of operational experience I bring to the table as a single environment. I would say I am familiar with a number of environments and yes, I am speaking from the perspective of the wide variety of environments and experience I have amassed by actively working in this industry for more than 25 years.

You say this as if it somehow makes me less than qualified to comment here. Nothing could be further from the truth.

Owen

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.afrinic.net/pipermail/rpd/attachments/20170728/41f9a8d8/attachment-0001.html>