[ppml] Policy Proposal 2004-3 Global Addresses for Private Ne twork Inter -Connectivity

[Paul Vixie]

i saw randy's comment...

> ...  private address space is a disease that
> should not be officially encouraged in any way.

...and i think i disagree.  the internet is a network of networks, and
if some of those networks prefer nonuniversal EIDs then that's autonomy
in action.  speaking as a rootop, i know that private/nonuniversal
address space is responsible for a huge share of my unanswerable queries,
and i wish we had better fences so that better fences could make better
neighbors.  but aside from the toxics leaching over the property line,
"private" address space is a strictly local, strictly autonomous issue,
and not a "disease" at all.

the second part of what randy said strikes me as right, though.  "private"
(to whom?) addresses ought not be officially encouraged.  i'm only saying,
these nonuniversal addresses should not be officiall discouraged, either.

background:

universality of EID's is quite expensive, both in overall internet
maintainability and by "encouraging" an economic model that favours
older and larger entities at the expense of newer and smaller ones.  we
cannot be seen kicking the crutches out from under enterprises who want
to limit their tithe, and avoid long term contracts with "the tiers",
and build their corner of "the internet" as a grid rather than pyramid.

even though i think it's insane that we use /48's for campuses and /64's
for LANs and that V6 will still have to reach swamp-equilibrium as part
of its success (if it succeeds), i have to say, i can imagine worse things.