[arin-ppml] ARIN-prop-127: Shared Transition Space for IPv4 Address Extension

[George Bonser]

> I'm afraid E has been so polluted by code restraints, there is no
> proper
> recovery of it. You're asking for a support nightmare.
> 
> Jack


I don't believe so.  So lets say Vendor C produces both a line of NAT
devices and a line of CPE that supports the use of E space.  Now lets
say other vendors in their newer gear (and software updates for their
older gear) also support E space.

Now you won't necessarily be able to renumber people out of their
current global space in all cases. But you can keep the problem from
getting worse by specifying the CPE for NEW connections that does
support it.  Anyone attempting to use such space internally will find
that their workstations and servers won't work with it.
 
What is going to happen with this scheme is everyone and their brother
are going to end up using this middle-4 /10 net for all sorts of
purposes.  To the extent that you can say you have no sympathy for
people who do that and eventually run into trouble, you can also say you
have no sympathy for people who have gone this long with v4 only stuff
in their network.

Mark my words, it will become defacto 1918 space within weeks after this
being accepted and people are going to start running into those
addresses in all sorts of places they aren't supposed to.

It is going to take an imaginative network engineer no more than about
15 minutes to realize they can reclaim all of their space they are
currently using for /30 or /31 WAN links even where they aren't doing
NAT at all and replace it with this address space.