[ppml] 2005-1 and/or Multi6

[Michael.Dillon at radianz.com]

> Can anyone please define "swamp"? What does the collection of PI
> prefixes differ in compared to the collection of PA aggregate prefixes
> other than probably prefix length?

My definition of "swamp" is an IP address range in which
many LONG prefixes (small blocks) are allocated for us in
a provider independent (PI) manner with little to no
possibility of aggregating those prefixes into shorter
ones.

In IPv6, this is bad because it means that more bits
will have to be consumed by the global routing table,
communicated in BGP announcements, and processed in
routing calculations. 

I do not believe that we will see the same improvements
in memory capacity, processor power, and circuit
capacity that we saw in IPv4's heyday because we are
now approaching real physical limits to such improvement.
At the same time, IPv6 increases the average prefix length
even if /32 is the longest prefix to be announced. Injecting
many /48 prefixes gives another 50% increase in prefix length
and this all comes at a time when more and more so-called 
end sites are moving towards PI addresses and multihoming.

We have to allocate PI addresses to these so-called 
end-sites because to do otherwise is restraint of trade.
However, it is unwise and imprudent to offer them /48
prefixes which are highly wasteful of global router capacity
when we could give them shorter /32 prefixes. In doing so
we are leveraging the vast oversupply of IPv6 address space
to help conserve the processor, memory, and circuit capacity
of network operators.

--Michael Dillon
P.S. I say "so-called" end sites because I believe that it
is untrue to view the Internet as something which has a center
and edges where the edges are populated by end-sites. Those
so-called end-sites are often internetworks in their own right.