Re: Resolving geo discussions

[Tim Chown <tjc@ecs.soton.ac.uk>]

On Mon, Apr 14, 2003 at 11:59:02AM -0400, J. Noel Chiappa wrote:
>     > From: Brian E Carpenter <brian@hursley.ibm.com>
> 
>     > There's no doubt that if you can persuade all the upstream ISPs to do
>     > the right thing, and persuade all the local downstream ISPs to do the
>     > right thing, the IX-based addressing model can help a lot (and can
>     > simply use PA space as if it was PI). But I don't think the economic
>     > forces involved will exercise the required persuasion.
> 
> As an aside, and also in tune with your comment, people might want to note
> that getting addresses from a local exchange (what I used to call a
> "cooperative exchange point") has a big advantage, which is that you're not
> tied to a single supplier; i.e. it's effectively number portability, as the
> exchange can switch members in the set of carriers providing service without
> anyone renumbering.

That is the plus, and what I understand Euro6IX is investigating.
 
> Of course, if you decide you're unhappy with the exchange, and want to leave,
> then you are stuck! :-)

Well, there are only so many exchanges in any big city :)   But if you
can arrange to stay with the same exchange, your chances of retaining
your address block from that exchange are higher.

> Also, this may lead ISP's to decide that they'd rather not support exchanges,
> since the customers are less locked-in (q.v. your comments about economic
> forces)! However, I'd imagine that if the exchange is big enough, and there
> are enough competing carriers, some carrier(s) will decide they'd rather have
> less-locked-in business, than no business at all. After all, plenty of markets
> operate without this level of lock-in.

Hopefully someone on the network modelling side of Euro6IX can comment.

Tim