On 27-mei-04, at 9:32, <john.loughney@nokia.com> wrote:
OK, so your point was:
1. The name or address value as it appears to the user. (FQDN, textual
representation of an IPv4 or IPv6 address, port number usually
implied.)
This isn't really a point... I was just listing the places where
names/addresses are used in some way in IP communication. This is one of
them.
WRT to Multi6, I think that this should be a non-goal - you seem to
be pre-supposing that the user may see the name.
Someone has to see something to begin with. After that there can be
invisible referrals.
I would imagine we have many possiblities, not restricted to:
1) URI, FQDN, etc. <-> multi6 name <-> IP address used (shim layer)
2) Multi6 address is in the form of some URI, FQDN, etc. (application
level address)
3) Multi6 address is in the form of an IP address (something like HIP)
So how does 1 compare to 3? Having a special multi6 name (apart from the
FQDN) that isn't in the form of an IPv6 address seems like asking for
trouble to me.
I think the "9 step program" illuminates the difference between 2 and 3:
Regular IP multi6=fqdn multi6=address
1. s=0, d=FQDNX unchanged unchanged
2. | s=A, d=X changes req. unchanged
3. | s=A, d=X depends changes req.
4. | s=A, d=X depends depends
--- i --- n --- --- t --- r --- a --- n --- s --- i --- t ---
5. | s=A, d=X depends depends
6. | s=A, d=X depends depends
7. | s=A, d=X depends changes req.
8. | s=A, d=X changes req. unchanged
9. s=FQDNA, d=X unchanged unchanged <-- (firewalling issues)
Referrals: changes req. depends
I'm not sure worry about this point is useful.
I'm not worried. :-)