On 30-mei-2005, at 22:23, Bound, Jim wrote:
Folks, the purpose of this thread is to define the purpose of the bits
for ND and addrconf not resolve how dhc works. We need to finish that
first ok.
The router is sending m and o bits now. What is their purpose and do
they work. If we change them it affects far more than dhc.
The thread to how to do this in dhc is most likely truly a dhc wg
effort
and not an ipv6 wg effort.
Again the m or o bit are to inform clients before they even
contemplate
dhc within an ND packet.
Do we need both bits but we need one for them for sure.
What I sense is that people who are interested in using DHCPv6 want
to be able to do DHCPv6 even if the routers don't explicitly say it's
ok, while others want to make sure the routers can instruct hosts to
NOT do DHCPv6.
So what about this:
M = 0, O = 0: hosts use their default mechanisms. There is no
assumption that a host will use DHCPv6, but there is also no
assumption that a host will NOT use DHCPv6.
M = 1, O = 0: hosts use full stateful DCHPv6 to obtain addresses and/
or other information to the degree that they can. Obviously there are
hosts that don't implement DHCPv6 or don't implement address
assignment, but the assumption is that DHCP will be used when possible.
M = 0, O = 1: hosts use some kind of stateless DHCPv6 to obtain other
information. DHCPv6 servers won't offer address information, and
hosts don't ask for it, and reject it when they get it anyway. The
exact DHCPv6 variation that should be used in this case will be
determined by the dhc wg asap, the idea being that in this case the
interaction and implementations are as light weight as possible.
M = 1, O = 1: hosts do NOT use DHCPv6, overriding any default behavior.
I think this accommodates all needs, with only the slight
inconvenience that the M = 1, O = 1 behavior is the exact opposite
from what we are supposed to have now.
I would be happy to receive yes/no comments (or more) in private
email and summarize those to the lists if people have an opinion but
don't want to join the discussion and not waste bandwidth with
monosyllabic posts.
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------