On 24 okt 2008, at 18:38, David W. Hankins wrote:
What if DHCPv6 servers do not exist? AFAIK, in this thread, the cost of unnecessary multicast DHCPv6 messages which may be prevented through M&O bits are being discussed.
That's a separate issue. Iljitsch is proposing that Multicasts must be reduced to as close to zero as possible, as a performance criteria.
Obviously multicasts are useful so I'm not saying we can't have any. But in the case above, where there is no DHCPv6 server so the client will be retransmitting multicast request, is clearly suboptimal, especially since we have a mechanism to avoid it.
If a client sends a DHCPv6 packet, it is because it requires some DHCPv6 configuration to use network services.
That's one of the problems with DHCP: the client doesn't know whether it needs configuration info, it's the server that knows whether clients need info. But it's the clients job to initiate the communication.
So the client is pretty much forced to run DHCP in case the server has something to tell it, unless we use a non-DHCP mechanism to avoid this when it's not necessary.
-------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
