[6bone] comments on draft-fink-6bone-phaseout-00

[Ronald van der Pol]

On Wed, Jan 22, 2003 at 18:58:14 -0800, Bob Fink wrote:

> >The IANA MUST reclaim the
> >   3FFE::/16 prefix upon the date specified in 2.0, and MUST make
> >   provisions to set it aside from any other uses for a period of at
> >   least two years after this date to minimize confusion with its
> >   current use for the 6bone (e.g., thus allowing production IPv6
> >   networks to filter out the use of the 3FFE::/16 prefix for a
> >   reasonable time after the 6bone phaseout).
> >
> >==> I'm not sure about the second MUST.  Perhaps a SHOULD would do?  For
> >example, consider if someone specified a locator,identifier separation
> >mechanisms which would use two IPv6 addresses.  Identifiers would be from
> >3000::/4 and the rest would be as before.  The above wording as I read it
> >would prevent the allocation of 3000::/4.
> 
> No, it only means 3ffe::/16 as it says, not anything shorter like 3000::/4. 
> I can add the prefix again in the wording if you think it makes a 
> difference.

I think what Pekka is saying is that reserving 3ffe::/16 prevents usage
of any prefix that includes 3ffe::/16, e.g. 3000::/4.

I think it is unlikely that we need a /4 for experiments in the next two
years (let's first come up with a plan for ID/LOC separation :-) There are
plenty of /16s in 2000::/4 or 3000::/4 that can be used for experiments.
Or we can use any of the other unassigned blocks.

But on the other hand a MUST is strong. Maybe SHOULD is better.

	rvdp