RE: PI space for very big sites [Re: stable addressing]

["marcelo bagnulo" <mbagnulo@ing.uc3m.es>]

Hi Pekka,

>
> On Wed, 21 Apr 2004, Brian E Carpenter wrote:
> > If it really is only 500 or 1000 enterprises world-wide, they
> will in the
> > end get PI space and it will be routed in the DFZ, by simple economics.
> >
> > The challenge is not so much there. It's
> >
> > a) to solve the problem for the few million companies that aren't as big
> > as Boeing but are too big to be connected in a simple way to a
> single ISP.
> > That's why multi6 is here.
> [...]
>
> Which is why I've argued that it might make sense to try to find a
> definition which could be used to give PI space to these very big
> enterprises, while disallowing to the rest.
>
> The only difficulty is drawing the line, but if someone could do
> that, we might be better off.

I agree that very big sites probably will get PI.
I also agree that the complex issue is to define the criteria used for
asigning.

But i think this is not a problem for this wg nor the IETF, imho this is a
problem for the RIRs

Regards, marcelo

>
> Still, in many cases PI-space might not be the optimal solution if the
> prefix would not be accepted as de-aggregated, if the enterprises
> wouldn't want to advertise the whol /32 in all the peering/upstream
> points (and handle the internal connectivity using more expensive
> links -- this was the case taken up by C. Huegen from Cisco).
>
> So, I'm not sure how much PI-space would *really* help because often
> that would require advertising more specifics for TE purposes as well,
> and that's not something we'd want to go for...
>
> --
> Pekka Savola                 "You each name yourselves king, yet the
> Netcore Oy                    kingdom bleeds."
> Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
>
>