RE: Regionally aggregatable address space for multihoming

[Iljitsch van Beijnum <iljitsch@muada.com>]

On Tue, 12 Jun 2001, RJ Atkinson wrote:

>         However, let me suggest that giving and disagreeing
> about examples aren't really the focus here, because the world 
> is a much larger place than the EU (or the US or AU for that 
> matter :-).

The point of the story is that a lot can change in five years.

>         The mere fact that in many places (not all, just many)
> local-loop charges are higher than long-distance charges means
> that the model we need must be economically sensible when the
> long-distance is less expensive than the local loop.

Long distance and local loop are really apples and oranges. From my living
room to New York is much, much more expensive than from my living room to the
other side of the city. However, from the Worldcom colocation building to New
York is probably cheaper than to the other side of the city. But from the
Worldcom colocation building to the Level3 colocation building is probably a
lot cheaper than from either to New York, and certainly than from Worldcom to
New York and then back to Level3.

And even if a connection over New York is cheaper, it is still possible to
exchange layer 3 local traffic over layer 1/2 long distance connections if
this keeps the routing table from exploding.

Iljitsch