IPv6 Thought Experiment

Mailman List Mirror (Read Only)
1

Dear list,
First of all, let me apologize if this post is not allowed by the list. To my best interpretation of the guidelines [1] it is allowed, but may be in a gray area due to rule #7.

I would like to propose the following thought experiment about IPv6, and I would like your opinion on what you believe would happen in such a case. Feel free to reply on or off list.

What if, globally, and starting at January 1st, 2020, someone (imagine a government or similar, but with global reach) imposed an IPv4 tax. For every IPv4 address on the Global Internet Routing Table, you had to pay a tax. Let’s assume that this can be imposed, must be paid, and cannot be avoided using some loophole. Let’s say that this tax would be $2, and it would double, every 3 or 6 months.

What do you think would happen? Would it be the only way to reach 100% IPv6 deployment, or even that wouldn’t be sufficient?

And for bonus points, consider the following: what if all certification bodies of equipment, for certifications like FCC’s or CE in Europe, for applications after Jan 1st 2023 would include a “MUST NOT support IPv4”…

What I am trying to understand is whether deploying IPv6 is a pure financial problem. If it is, in this scenario, it would very very soon become much more pricey to not deploy it.

I know there are a lot of gaps in this, for example who imposes this, what is the “Global Internet Routing Table”, etc. but let’s try to see around them, to the core idea behind them.

Thanks,
Antonis

2

To clarify that further, this would be a monthly tax. So $2 / month.

3

Antonios,

It’s certainly financial but it’s not just companies being cheap. For example for smaller companies with a limited staff and small margins. They may want to have v6 everywhere but lack the resources to do it. It would for certain speed up the process but there would be collateral damage in the process.

4

It appears in your thought experiment, a stick is dressed up like a carrot.

I’m not a fan of deploying purely punitive strategies to promote adoption; technologies should stand on their own and be able to convince the potential users based on their merit, not based on penalties.

5

Let me clarify that I 100% agree with both Job and Dovid. It is indeed a terrible idea. And not everyone is even convinced IPv6 is the right next step. So it’s obviously wrong to push people towards where someone thinks, even if it’s the majority.

I just had a hunch that even then we would still not see IPv6 adoption.. So there must be many more problems that we overlook. I just wanted to hear what other people think on the matter.

6

By virtue of depletion at the RIRs, there’s effectively already a one-time IPv4 tax, the cost of procuring the addresses. This has indeed increased over time, and eventually we will reach a point where for many organizations acquiring IPv4 address space is not realistic either because they cannot afford it or (if you look at someone like AWS/Azure/etc who blow through lots of addresses) just won’t be able to acquire the scale they need. This is happening on its own.

IPv6 deployment is happening, albeit slowly. Mobile providers are increasingly using IPv6 traffic to avoid having to push more CGNAT gear, consumer and small business ISPs are getting on board bit by bit, and while there may have been some point a bunch of years ago in looking into ways to speed adoption prior to the RIR depletion situation that we’re now faced with, I’m not sure there’s any meaningful benefit to trying to artificially push things forward at this point.

7

And for bonus points, consider the following: what if all certification bodies of equipment, for certifications like FCC’s or CE in Europe, for applications after Jan 1st 2023 would include a “MUST NOT support IPv4”…

I think a good start would be: “MUST support IPv6”!

8

For a small organization with limited staff and small margins, I’m curious where the actual burden in supporting IPv6 lies. In my experience, it’s not any more costly than deploying IPv4 is (and really, less so over the past couple of years since you can get IPv6 RIR allocations while adding IPv4 capacity means shelling out thousands or tens of thousands of capex dollars.) I’ve never had an IX or transit provider or anyone else charge me more because I’m running IPv6 in addition to my IPv4, and any gear that doesn’t support IPv6 at this point is likely old enough to be EoL and requiring replacement due to potential (major, very costly) security issues anyhow.

9

A few thoughts:

  1. What global organization has the ability to impose a tax on any nation’s citizens?

  2. Do you not see an issue with making everyone worldwide get rid of every device that supports v4? Kind of a burden for a developing country, no? Also, a bit of an e-waste problem I would think.

  3. Do you think that any organization with the power to tax some Internet usage (like v6) will stop there and not figure a way of continuing the cash flow forever?

  4. The FCC and other standardization organizations often have statutory authority to manage things like spectrum management and consumer safety. What would be their authority to mandate v6 usage?

  5. Why not just get carriers to make v4 service an optional extra just like static address requests? There is no reason to empower government more than they already are. Simple economic pressure would work.

  6. Why is your issue more important than any other so-called global issue like carbon taxes, endangered species, human trafficking, etc? Do you want to go to a world government to encourage adoption of IPv6? Why should anyone care about that other than us engineers working under the hood

  7. If someone like say Botswana says we are not paying your tax, do you intend to send in UN Peacekeeping Forces to collect the money owed? Are we going to war with North Korean if they won’t let us check their routers for the presence of v4 addresses?

  8. What is the economic or social reasoning behind obsoleting ipV6? Is this really an existential global issue or are you just inconvenienced by dealing with both address families? While we think it a big deal here on NANOG, do you really think that the public sees that issue somewhere in their top 20 priorities? I doubt it.

  9. Some world government enforcing global network standard migrations? What could possibly go wrong there J. Do permanent UN Security Council members retain the right to veto these standards?

  10. I think at one time the US Government demanded POSIX compliance for all of their systems. That did not even work on the scale of the US Government managing their own systems. Why would this work any better? Governments are notoriously bad at managing their own IT systems, I don’t think we want them managing all of ours as well.

Steven Naslund

Chicago IL

10

In article <5DCAE7A8-1D33-4EA2-BBB1-7A3E8132D55B@gmail.com> you write:

What do you think would happen? Would it be the only way to reach 100% IPv6 deployment, or even that wouldn’t be sufficient?

If you have to impose an artificial tax to force people to use IPv6,
you've clearly admitted that IPv6 is a failure and can't stand on its
own merits. Should this happen, I'd expect massive use of CGN to hide
entire networks behind a single IPv4 address, and a mass exodus of
hosting business to other places which are not so stupid. Mobile networks
would be less affected because many of them are IPv6 internally already.

What I am trying to understand is whether deploying IPv6 is a pure financial problem.

To some degree, anything is a financial problem. How about if I
charge you a hundred dollars for every packet you send using IP rather
than CLNS and CLNP and a thousand dollars for every virtual circuit
using TCP rather than X.25?

11

It's certainly financial but it's not just companies being cheap. For example for smaller companies with a limited staff and small margins. They may want to have v6 everywhere but lack the resources to do it. It would for certain speed up the process but there would be collateral damage in the process.

Here is the question being dealt with in the corporate environment. Why should I prioritize moving everything to IPv6 now instead of my other zillion IT projects that actually are visible to my customers and business users? It is simply, almost always, a cost benefit question. I would have to convince the company that it is in their financial best interest to go that route. I think over time the migration happens organically as more people are familiar with v6 and all the equipment and setup schemes start using v6 as the default. I would be hard pressed to come up with a reason for a hard deadline. Making life easier for NANOG engineers is not high on most corporate priority lists :slight_smile:

Steven Naslund
Chicago IL

12

I understand, but I think there’s something else here.. If we keep deploying IPv6 at this rate, we will have it in X years. If such a policy was enforced, it would *accelerate* the transition, not force it. It would kind of force it in a way that companies that can’t comply would maybe seize to exist, but overall it would accelerate the IPv6 adoption. This is the main thing I was trying to “achieve” here. And the question is, if instead of reaching 100% deployment in X years, we reached it in X/2 or X/10 or X/100, would we be ready? Would the RIRs be ready? Would the vendors be ready? Would the equipment be ready? Could we sustain the increase of the routing tables?

Maybe all of us got kinda lazy because it’s moving at a so slow pace.. I don’t know, maybe we didn’t, and we just have so many problems that we solve them at literally the last possible moment (or a bit after that).

Antonis

13

Dear list,
First of all, let me apologize if this post is not allowed by the
list. To my best interpretation of the guidelines [1] it is allowed, but
may be in a gray area due to rule #7.

I would like to propose the following thought experiment about IPv6,
and I would like your opinion on what you believe would happen in such a
case. Feel free to reply on or off list.

What if, globally, and starting at January 1st, 2020, someone
(imagine a government or similar, but with global reach) imposed an
IPv4 tax. For every IPv4 address on the Global Internet Routing
Table, you had to pay a tax. Let’s assume that this can be imposed,
must be paid, and cannot be avoided using some loophole. Let’s say
that this tax would be $2, and it would double, every 3 or 6 months.

Who exactly would be paying this tax? The IPv4 address "owner"? The
SWIP? The end user who gets IPv4 via DHCP from his provider?

Tax paid to whom?

What do you think would happen? Would it be the only way to reach
100% IPv6 deployment, or even that wouldn’t be sufficient?

Well, a lot of money would change hands. Somebody would be enriched by
the tax revenues.

And for bonus points, consider the following: what if all
certification bodies of equipment, for certifications like FCC’s or CE
in Europe, for applications after Jan 1st 2023 would include a “MUST
NOT support IPv4”..

So how would that affect users trying to access IPv4 resources?

What I am trying to understand is whether deploying IPv6 is a pure
financial problem. If it is, in this scenario, it would very very soon
become much more pricey to not deploy it.

First, there are equipment issues -- not all gear "plays nice" with
IPv6, especially older gear still in use. There is a capital cost
associated with upgrading gear, and not all organizations and people can
afford the hit.

There are policies in place, beyond the RFCs, by companies and
governments that would need to be updated, and the tax you suggest
doesn't even begin to attack the problems.

I know there are a lot of gaps in this, for example who imposes this,
what is the "Global Internet Routing Table", etc. but let’s try to
see around them, to the core idea behind them.

Has BCP-38 been updated to include IPv6?
https://tools.ietf.org/html/bcp38

All the examples are IPv4. Additionally, one of the reference is this:
Baker, F., "Requirements for IP Version 4 Routers", RFC 1812, June 1995.

If people are serious about IPv6, isn't it time to update the Best
Practices documents, particularly BCP-38 et al, to address IPv6 as well
as IPv4?

14

In article <CAHdm834jbwky2sPpH6HmJoYu=Rcjz0Hb1bCq2zy1hsdYOSN9sQ@mail.gmail.com> you write:

For a small organization with limited staff and small margins, I'm curious
where the actual burden in supporting IPv6 lies. In my experience, it's not
any more costly than deploying IPv4 is ...

Right, but that means it doubles your deployment costs since IPv4
isn't going away any time soon. First you have to get IPv6 into your
network, directly or through a tunnel (thanks, HE.) Then you have to
assign IPv6 addresses to every device that has a name, put that in
your DNS and configure the devices, either by whatever means the
device has (typically a web control panel) or maybe by a DHCP entry,
if the device can be persuaded to use DHCP rather than SLAAC. In
many cases, notably web servers, you need yet more configuration to
connect each v6 address with whatever service the v6 adddress is
supposed to provide.

Then you have to set up firewall rules to match your v4 firewall rules.

Then you spin it all up, and you have to check that every device
actually does respond on its IPv6 address, and that it acts reasonably
to mixed v4 and v6 requests (so-called happy eyeballs.)

None of this is impossible, I've done it all, but I've also often
asked myself what exactly is the benefit of doing all this. On my
home network the v4 stuff is behind a NAT so v6 allows me access
to devices from the outside (carefully managed with the firewall)
but on my hosted servers which have v4 addresses for everything, meh.

15

Interesting idea. Let's say it started off at $2 / month and doubled every 3 months. At the end of month 12, it would be $32/month. After 5 years, we'd be talking about just over $2 million per IP address per month, i.e. a little over half a billion dollars per /24. In 10 years, that would increase to 562 trillion dollars per month for a /24.

Soon, you'd be talking about real money.

Please let me know if you wish to push ahead with this idea and I'll humbly offer to act as middle-man for taxation for a very simple and modest 1% of all transaction fees, or 0.94% if you can guarantee an exclusive deal. Serious replies only please.

Nick

16

What happens when v4 is gone? Surely you won’t let it end there - After all, If you have the ability and infrastructure to do this, why not tax IPv6 too? This would cut down on the number of “undesirables" on the internet by pricing it out of the reach of all but the largest megacorporations. Eventually we can reduce the internet to a few dozen authorized parties in each region and we’ll only need enough IP addresses for those. I can imagine a number of governments around the world would be very interested in this.

17

In article <CAHdm834jbwky2sPpH6HmJoYu=Rcjz0Hb1bCq2zy1hsdYOSN9sQ@mail.gmail.com> you write:

For a small organization with limited staff and small margins, I’m curious
where the actual burden in supporting IPv6 lies. In my experience, it’s not
any more costly than deploying IPv4 is …

Right, but that means it doubles your deployment costs since IPv4
isn’t going away any time soon. First you have to get IPv6 into your

I’d strongly disagree that it anywhere near doubles costs. Ultimately you’re buying hardware X and it’s going to cost whatever it costs. So what more do you really need to do to support IPv6? Well, let’s say you’re using OSPF. This means you’ll also need to use OSPFv3, but that’s not hard because your OSPFv3 configs are going to basically mirror your OSPF configs. You’ll need to run IPv6 over iBGP, perhaps, and over eBGP to your peers and transits, but that’s just another set of addresses bound to interfaces, sessions that mirror the IPv4 ones, and policy rules/filters. If you’re doing super heavy TE, then the filter configs might take some effort, but if we’re talking about smaller shops, doing heavy TE is unlikely. At that point, you just add a v6 address to your layer3 interfaces and you’re good to go for the network side.

Most of the time you spend configuring things won’t be v4 or v6 specific, and the v4 specific configurations can be copy/pasted with a quick string swap to support v6 in a lot of cases.

network, directly or through a tunnel (thanks, HE.) Then you have to
assign IPv6 addresses to every device that has a name, put that in
your DNS and configure the devices, either by whatever means the
device has (typically a web control panel) or maybe by a DHCP entry,

But once you have the basics down - your layer3 gateways, routing protocols, etc - this process of getting the hosts on board can take as long as you’d like. The only deadlines are ones you impose.

if the device can be persuaded to use DHCP rather than SLAAC. In
many cases, notably web servers, you need yet more configuration to
connect each v6 address with whatever service the v6 adddress is
supposed to provide.

I’ve done this plenty of times and never found it to be particularly difficult or time-consuming. If you have a lot of systems, hopefully you have some sort of automation or configuration management which will allow you to test and deploy to production in at least a less-manual fashion. If you don’t have a lot of systems, then you can just iterate through them and take 10-15 minutes each to get a v6 address bound and firewall rules updated.

Then you have to set up firewall rules to match your v4 firewall rules.

Which usually means copy/pasting and a quick string swap.

Then you spin it all up, and you have to check that every device
actually does respond on its IPv6 address, and that it acts reasonably
to mixed v4 and v6 requests (so-called happy eyeballs.)

Or just throw it in the monitoring system you already have, but again, this can all be part of a process that happens over as long a course of time as you need it to.

None of this is impossible, I’ve done it all, but I’ve also often
asked myself what exactly is the benefit of doing all this. On my
home network the v4 stuff is behind a NAT so v6 allows me access
to devices from the outside (carefully managed with the firewall)
but on my hosted servers which have v4 addresses for everything, meh.

I think ultimately the perception of the work required to deploy IPv6 is a much greater hurdle to IPv6 adoption than the actual work required to deploy IPv6.

18

I suspect that even if there was an entity with the reach to impose such a tax, people will resort to deploying CGN more, to hide their IPv4 usage to the extent possible. That’s time, money, and effort taken away from moving to IPv6.

You might also find that many taxed organizations will simply ignore the tax or refuse to pay it, under the assumption that the taxing entity doesn’t have standing to impose such taxes. Someone from Russia is likely to take a tax notice from, say, some agency in the USA and toss it in the circular file :slight_smile:

As others have said, threatening the Internet community with punitive action is a sure way to discourage people from adopting IPv6. While the pace of adoption might not be acceptable to some, everyone has to move at their own pace, or vote with their wallets where possible.

Thank you
jms

19

I'm describing my actual experience, so we'll have to disagree here.

Regards,
John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
Please consider the environment before reading this e-mail. https://jl.ly

20

In my experience, the biggest hurdle to installing a pure IPv6 has nothing to do with network gear or network engineers. That stuff I expect to support v6. This biggest hurdle is the dumb stuff like machinery interfaces, surveillance devices, the must have IP interface on such and such of an obsolete appliance, etc. The dumb legacy app that supports the ancient obsolete pen plotter that we must keep forever, etc.

The next largest hurdle is trying to explain to your server guys that you are going to go with all dynamically assigned addressing now and explaining to your system admin that can’t get a net mask in v4 figured out, how to configure their systems for IPv6. There are a lot of people in the IT industry who are not nearly ready for v6. In large enterprise networks, there is lots of East/West communications between systems and that is very difficult to transition through a dual stack process without tripping over a bug or serious incident.

It is really hard to look at cost difference in v6/v4 but there will be a definite learning curve with the associate oops moments and re-education that all costs time/money/downtime. The simply reality is that there are more IT people that understand v4 and do not understand v6 yet.

Steven Naslund

Chicago IL