[arin-ppml] End non-public IPv4 assignments?
bicknell at ufp.org
Mon Jan 24 12:26:59 EST 2011
In a message written on Mon, Jan 24, 2011 at 12:08:01PM -0500, William Herrin wrote:
> Playing devil's advocate here: if blocks from these networks will
> never appear on the Internet, why isn't the entire 32-bit address
> space available to them? What's wrong with overlap between addressing
> on the public Internet and addressing on strictly private networks
> like financial exchanges, check-clearing networks, electric smartmeter
> networks or DoD's SIPRnet? More to the point: what's wrong enough that
> it justifies removing those addresses from use on the largest TCP/IP
> network (the Internet) where there's a critical shortage?
The enterprise networks all connect to the public internet as well.
If the exchange network used the same range as say, Facebook then
the enterprrise would be unable to reach those web sites.
So while the actual exchange is not connected, everything that connects
to them is connected, thus the need to be unique. It's similar to the
Internet 2 backbone issue someone else described.
In diagram form:
Private Exchange Network
| | | |
Ent 1 Ent 2 Ent 3 Ent 4
| | | |
Public Internet Connectivity
None of the enterprises advertises the private exchange to the
public internet, so if you're on the public internet you can't see
it at all. However, they all connect (reason it must be unique)
and they also all connect to the public Internet (reason it must
be gobally unique).
I think from the description, replace my private exchange with
Internet 2's backbone and you have the same sort of situation over
Could either one be architected a different way, probably. However
I'm not going to wade down that slippery slope right now. We could
say this doesn't pass some hypothetical bar for "good use of
addresses", but that opens the can of worms of what is a good use?
Maybe dialup should now be denied. How about no more addresses for
porn hosters? Social networks are clearly just a waste of time,
so we should deny addresses to those folks.
We're out of IPv4. Even if we denied all non-connected addresses I
think you're going to "save" a few /24's over the next year. It's not
enough to even extend IPv4's lifetime by a few minutes.
Leo Bicknell - bicknell at ufp.org - CCIE 3440
PGP keys at http://www.ufp.org/~bicknell/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 826 bytes
Desc: not available
More information about the ARIN-PPML