Le 21/10/2012 02:45, Hesham Soliman a écrit :
The obvious conclusion to this argument is that a *lot* of DHCP
functionality will be duplicated in ND. Is this where we want to
go?
I'm coming from the DHCP side of the argument. In my world DHCP is
needed because it gives you a single place to handle dynamic
address allocation, *and* it ties in with all sorts of support &
backend systems.
=> Agreed. I'm not coming from a DHCP side of the argument but I see
the only reason for having this feature is that some people think
adding/implementing this in ND is more convenient than adding a DHCP
agent/relay. I don't agree. But even if that's the truth, it seems to
be six of one and half dozen of the other. No compelling reason
AFAICS.
There is already ND on MR-LV and MR-IV. And it's not enough to do
Prefix Delegation. We add a new protocol DHCP-PD but that is not enough
per se, it needs ND.
I wonder why do we always have to use two protocols when one should be
sufficient.
Alex
Hesham
I am against adding lots of new ND functionality until we have
DHCPv6 that is considerably more feature complete. Some of this is
probably coming (client MAC address), some of it is still being
opposed for mostly religious reasons (e.g. running DHCP without
RA).
Steinar Haug, AS 2116
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org Administrative Requests:
https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
--------------------------------------------------------------------
IETF IPv6 working group mailing list ipv6@ietf.org Administrative
Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
