[ppml] Proposed Policy: PI assignments for V6 (and v6 fees)

Michael.Dillon at radianz.com Michael.Dillon at radianz.com
Tue Dec 7 05:44:01 EST 2004

> i don't think it'll work like you're saying.  let's say ULA goes into
> general use and that a number of enterprises who are all customers of a
> set of ISP's decide that rather than tunnelling, they'd like their
> shared set of ISP's to "just route the ULA prefixes please."  by
> enterprise i'm talking Ford and GM and Daimler and their rust-belt 
> network, or perhaps Wal-Mart and its offshore vendor network.  we know
> what the ISP's will say: "yes, ma'am, and would you like fries with 

This sounds a lot like what the European auto industry
does with Deutsche Telekom, France Telecom, Telefonica
and Infonet. http://www.enxo.com
The original design for the ENX was to have these ISPs 
routing their traffic across separate peering connections 
from their other IP traffic. As far as I know it still
works that way.

The way to stop this is to create ULAs and also
create geographically agregatable addresses. Then
the vast majority of users who might otherwise have
chosen the ULA peering route will instead choose to
use geographically agregatable addresses. Traffic destined
to Ford's plant in Sacramento will always head towards California
regardless of which provider network it travels on. Do you
think Ford really cares if the traffic bounces off of 
San Francisco before reaching Sacramento?

> instead,
> let's just recognize that a far larger segment of the enterprise 
> needs PI than was previously believed, 

I strongly agree with this one. And I think we need a serious
assessment of geographic addressing before we decide how to 
solve this. By serious, I mean some hard numbers that show us
things like the number of households in every LATA in the USA
as well as every Canadian province. In IPv6, the household is
the basic unit of measure, better known as the /48. If we
build a geographical addressing scheme that will handle the current
number of households plus a bit, then we have something that will
last for many years. And when it starts getting tight in one region
or other, we can simply allocate a new continental agreggate because
the impact on the global routing table will be minimal. We should
be able to get this right with only these two allocations so that
it will last the entire lifetime of IPv6 as we know it.

> and transform all ULA arguments to
> date into "PI reform" arguments, and then restore "site local" 
> which was valuable precisely because it was dangerous and self-limiting.

Personally, I'm not that bothered by ULA addressing. And I
expect that any address range set aside for site local will end
up being used as ULA addresses because the random prefix idea is
too good to go away. In fact, if ULAs don't make it through the
IETF, then people will do this anyway.

The box is open, we now have to deal with it.

--Michael Dillon

More information about the ARIN-PPML mailing list