<snip>
=> I disagree: the problem still exists but in a different form: to make
a network map/topology you need to bind link-local and global addresses.
<end snip>
Not sure. Why do you assume that the global address tells you anything about
location? It would be better to have something that maps identities of devices
and at his turn potentially maps it into some kind of topologies.
It would be interesting however to have awareness of who is attached
and with which addresses (a use example would be for ISP billing purposes).
There are techniques out there that allow this to be done, but i doubt if
we need
a protocol for this purpose as such.
Groetjes,
G/
At 16:50 2/08/2005 +0200, Francis Dupont wrote:
In your previous mail you wrote:
> PS: note both proposals are based on ICMP name lookups in order to
> get rid of the basic issue: link-local addresses are not useful when
> you are not on the link.
Sure, and this problem wouldn't exist with the multicast approach, as
the source address that the hosts would pick in the responses would be
global/ULA.
=> I disagree: the problem still exists but in a different form: to make
a network map/topology you need to bind link-local and global addresses.
Regards
[EMAIL PROTECTED]
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
