[arin-ppml] how to get a v6 /32 of v4 address space
hostmaster at uneedus.com
hostmaster at uneedus.com
Sun May 19 13:32:49 EDT 2019
Nor to me. A single /64 IPv6 subnet which is the standard LAN assignment
has 2^32 more addresses than the TOTAL IPv4 space. While unused IPv4 space
is being sold and pressed into use, and other space belonging to others
like 188.8.131.52/8 and 184.108.40.206/8 is being used for private space, eventually
the sheer growth of the Internet and networks will get to the point that
even this will not be enough.
Seventeen /8's were assigned to RIR's in 2010, the year before IANA
exhaustion happened. That was 1/13 of the total space available.
I have no doubt had space existed in the last eight full years since
exhaustion that at least another 136 /8's would have been consumed at that
same seventeen per year rate, maybe more.
While those with legacy elements in their network that prevents these
operators from easily adopting IPv6, my guess these operators are but a
very small fraction of the total, and likely already has enough IPv4
already in use. I understand legacy, as I still have IPX netware on my
network for a software program for Adult Basic Education. This is because
the current COTS software goes no lower than 8th grade, and we still have
a need to go below that grade. Workstations are booted off of floppies
(more legacy) to access the local Netware Server (controlled by KVM)
instead of the more modern stuff. Might make the workstations dual boot in
the next cycle, as the next generation of machines to be passed down to
the learning labs do not have floppies. Also, voicemail computers running
DOS. We do use ip based KVM switches (less the "M", unused in DOS) for
remote control, and dos based packet drivers and old ISA network cards for
remote drive access for backup. Even with all this old stuff, we have
been doing IPv6 since 2007 due to a Federal requirement to do so.
The "IPv4 Market" is not sustainable in the longer run, and my guess is
that in another 8 years that the price for IPv4 address will finally rise
to the point that it will be cheaper for most operators to adopt IPv6,
rather than pay the higher market cost to buy IPv4 from someone else. Of
course, some early adopters of IPv6 will take advantage of this increase
in market rate by selling off most of their unused IPv4 addresses. I also
believe that starting first with residential customers, the assigning of a
global IPv4 address will become an extra cost option, and maybe some that
are really bold will make access to IPv4 an extra cost service, making
IPv6 only service their lowest cost tier.
Of course, when we reach such a tipping point and the majority of
providers are pushing bits mostly using IPv6, market forces will then
start taking the price of IPv4 back downward. At the same time some
networks might limit or even eliminate support for IPv4 completely. Some
like Facebook are already completely IPv6 except at the edge.
There are still those that seem to be hanging on for some other protocol
other than IPv6. The reality is that development of IPv6 took over ten
years, and there will not be any time to develop something else.
While IPv6 is not perfect, it is at least as useable as IPv4 and been
available in most OS's and routers since 2000 or so. If one wants to hack
their OS in order to use 127/8 and 240/4 just to gain a few addresses that
is just a fraction of a normal IPv6 single LAN allocation of /64 go ahead.
However, it will still be likely that you will still at some point be
forced to use IPv6 when networks start turning down IPv4 support and your
network needs access to these other networks.
Paradise On Line Inc.
On Sun, 19 May 2019, Cynthia RevstrÃ¶m wrote:
> I have no clue what your point is but an IPv6 /32 is 2^96 IP addresses. The
> total possible IPv4 address space is 2^32
> So your point doesn't make much sense to me.
> - Cynthia
> On Sun, May 19, 2019 at 5:54 AM william manning <chinese.apricot at gmail.com>
>> ok, so you don't like the "use 127.0.0.0/8" proposal. fine.
>> RFC 1918 space is too small. fine.
>> IPv6 is too hard. fine.
>> Shortly after discussions started on RF 1918, I proposed the following:
>> Since NAT exists, direct peering on a global scale will be fairly
>> restrictive, one should consider inverting RFC 1918. Use those addresses
>> strictly and only for global interconnection/peering.
>> This would free up all other IPv4 space to sit behind your NAT and usable
>> in your enterprise networks. Thats almost an entire IPv6 /32 of space for
>> everyone, without having to migrate to IPv6.
>> Problem solved.
>> Your welcome.
>> You are receiving this message because you are subscribed to
>> the ARIN Public Policy Mailing List (ARIN-PPML at arin.net).
>> Unsubscribe or manage your mailing list subscription at:
>> Please contact info at arin.net if you experience any issues.
More information about the ARIN-PPML