[arin-ppml] Use of "reserved" address space.

[James Hess]

On Sat, Jun 26, 2010 at 11:21 PM, Joe Maimon <jmaimon at chl.com> wrote:
> What is wrong with an approach of "All of the above"?
> Anyways, strong odds suggest that removing restrictions on reserved space is
> a much simpler code change than including another network stack, in any OS
> and in any firmware.

This is definitely true.   Removing an arbitrary restriction on some
IPs is much simpler a change than adding an extra IP stack.  And it
should require at most a minor update to the standards, to adjust the
reserved blocks  to indicate host and router software should support
unicast for all those blocks from now on.

As far as I can tell , the internet is not on course for a
well-implemented  transition to IPv6.
It is on a collision course with IP exhaustion,  and the fallout of
the crash is NAT hell.


Even if a decision will not be immediately made to offer to allocate
address space from those blocks.
The availability of those  currently reserved blocks could definitely
be useful in the future,  every effort should be made as soon as
possible   to  encourage software being made now and in the future to
support those IPs....  provisions should be made immediately,  to
require all internet hosts to be capable of receive and send packets,
and be configured to use any parts of those special blocks that are
not   actually  used and can be unreserved.


You can implement IPv6 on your network, but you cannot force everyone
you want to talk to to do so at this point. The cost of getting other
people to use IPv6 would be very high.  The cost of deploying IPv6 on
your own network is irrelevent, once you have IPv6 networking,  it's
everyone else's network that matters.


The cost of asking other networks to patch their routers to fix
connectivity to formerly-reserved blocks is nil.
You can even report to them  "your equipment has this issue, bug,
please do something about it"..


Rather than having to say "We don't have IPv4 connectivity, will you
please setup an IPv6 network, so we can talk to you?"      Which is
more likely to get "No" as a response than  "Ok.. we'll see if we can
fix that"

The real cost of deploying IPv6 is not just a code change to
equipment, it is the IP network re-design,  planning, etc, required.

When the planning and design required to have a V6 network has not
been done,  the less costly alternative to V6 is  probably  NAT hell.

--
-JH