[c-nsp] IPv6 subnets for point-to-point links

[bmanning at vacation.karoshi.com]

On Sun, May 08, 2005 at 02:42:36PM +0200, Gert Doering wrote:
> Hi,
> 
> On Sun, May 08, 2005 at 12:25:05PM +0000, bmanning at vacation.karoshi.com wrote:
> > 	geoffs 0.94 work is interesting, but
> > 	neglects the savings/longevity of "violating"
> > 	the /64 restriction.
> 
> So how much savings (in terms of /48) do you estimate if all networks
> world wide change from using /64s for all transfer networks to /112.../126 
> (depending on the number of nodes)?

	i've not done the work.  geoff is wrapping up his H/D ratios
	@ 0.94 percent and it intimates changing that value by 3 bits gives
	us several more years before we run out.

	no rigourous analysis has been done on longevity of the space
	if we move to the right.

> Changing the /64 recommendation for "normal" LAN segments (with autoconfig
> etc.) is something I'd see quite some value in - more wiggle room to the
> front - but as has been said before "this is going to be hard and painful".

	one of the lemas used in early IPng discussions was to -NEVER-
	again get stuck in a fixed boundary condition wrt hardware.
	vendors who have a hardcoded or fixed boundary state on the
	/64 'split' have made a shortsighted error, imho.

--bill

> 
> gert
> 
> -- 
> USENET is *not* the non-clickable part of WWW!
>                                                            //www.muc.de/~gert/
> Gert Doering - Munich, Germany                             gert at greenie.muc.de
> fax: +49-89-35655025                        gert at net.informatik.tu-muenchen.de