% title = "Timetable for IPv4 to Deprecated and then Historical Status" % abbrev = "shutdown-IPv4" % category = "info" % ipr="trust200902" % docName = "draft-latour-shutdown-ipv4.txt" % area = "Internet Area" % workgroup = "Sunsetting IPv4" % keyword = ["sunset4"] % % date = 2016-02-05T00:00:00Z % % [[author]] % fullname = "Jacques Latour" % initials = "J." % surname = "Latour" % organization="CIRA" % [author.address] % street="Ottawa ,ON" % email="jacques.latour@cira.ca" % % [[author]] % fullname = "Owen DeLong" % initials = "O." % surname = "DeLong" % organization="DeLong Consulting" % [author.address] % street="" % email="owen@delong.com" %
.# Abstract IPv4 is unsustainable on the global internet. The increasing costs and declining performance of IPv4 based systems due to address overload are an unsustainable burden that is not only costly in terms of operational money, but also in placing artificial constraints on further development and by reducing the quality of the user experience.
IPv4 will die a natural death, but waiting for it to do so will take many more years than is currently desirable. Therefore, this I-D proposes to set a flag date to declare IPv4 no longer be supported as a globally routed protocol on the internet. The authors recognize that pockets of IPv4 for various specialized applications will endure for years and possibly even decades to come. However, at the time of writing, IPv4 is the current “lingua franca” of the internet. We propose to declare a date certain by which IPv6 will become the “lingua franca”.
{mainmatter}
Why is this needed ?
IPv4 is over 30 years old!
Dual-stack is the way to go for a while, but eventually the depletion of IPv4 addresses will cause IPv6 only infrastructure and services to segment the Internet, those on IPv6 only and those IPv4 only. If the majority of the growth on the Internet will be IPv6, we'll be left with small and isolated IPv4 networks. It's time to look at those networks now and see what can be done today to ensure they are part of the IPv6 infrastructure. Certainly the commercial entities will not support expensive IPv4 support too long if there are only a few isolated IPv4 only subnets.
Dual stack is an interim solution to a real problem, dual-stack has a cost, it's expensive to operate and doubles the security exposure. Dual-stack is temporary.
The dates specified are suggestion only and need to be agreed on, they are to provoke a reaction, to think about the impacts of shutting down IPv4, to think of the global actions we need to take in order to have an IPv6 only Internet.
There are many reasons to do this, but the most obvious one is that progress on any task without a specific deadline tends to proceed at a slower pace than one with a deadline. In this case, the lack of a deadline has lead o a 20+ year procrastination in deployment of IPv6. We have been suffering the consequences of IPv4 address shortage since the first deployment of NAT for address conservation. This has gone on for so long that there are those who have moved beyond acceptance of these limitations to a form of Stockholm syndrom-like support for NAT. Even with the various address sharing technologies and their attendant tradeoffs, IPv4 is strained and we are seeing increasing costs in operations, in address acquisition, and in various other aspects of maintaining an IPv4 network. In short, with IPv4, we are continually paying more to continually receive a less beneficial network.
By setting a date certain for the deprecation and eventual historical status of IPv4, the IETF can send a clear signal to developers, operators, administrators, and even the financial markets that this change is vital and necessary to the sustainable operation of the internet and that plans are in place for achieving this goal.
In terms if timelines, we propose to deprecate IPv4 in 2022/02/01 and move the status to historical on 2024/04/04.
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [@RFC2119].
The primary goal of this I-D is to define key dates and document a timeline and action plan that for the discontinuation of the IPv4 protocol on the global Internet. The dates need to be aggressive enough to avoid further deployment procrastination.
The dates need to allow sufficient time to identity potential IPv4 only networks (those who can't afford to upgrade their IPv4 only network equipment) and plan for upgrades.
The target date to have an IPv6 only Internet is April 4, 2024.
Yes, IPv4 can exist inside Internal networks, IPv4 can be tunneled over IPv6 on the global Internet. This I-D only addresses IPv4 on the global Internet.
A phased approach is recommended over a few years to gradually remove the dependence on IPv4.
We should start deprecating IPv4 names and finish with removing all global Internet IPv4 routes.
Remove IPv4 NS records Remove IPv4 glue records Remove IPv4 registry access, EPP access
This phase is schedule for 2022/02/01
TLD Domains names are IPv6 only The majority of Internet traffic is over IPv6
ARPA zone management, remove all IPv4 related zones; Remove in-addr.arpa Remove IPv4 in as112.arpa Remove any IANA IPv4 references
This phase is schedule for 2023/02/01
Find pockets of IPv4 that are lagging on the migration Reverse DNS is IPv6 only The vast majority of Internet traffic is over IPv6
This is the last step in shutting down IPv4, remove IPv4 routes from all AS operator.
This phase is schedule for 2024/02/01
Smaller core Internet routers.
Reduced operating costs.
Big Party!
That's it for IPv4. List all the thing to make IPv4 Historical
This phase is schedule for 2024/04/04
Business as usual
This will hopefully make the Internet more scalable, secure and cost effective to operate and manage.
Lots, TBD
This protocol is designed for machine to machine communications
{backmatter}
First rough version