As Jinmei-san pointed out, RFC 3736 provides implementation guidelines for servers, relay agents and clients.
After reading this thread, I think the text currently in RFC2461bis is headed in the right direction; that is, it explicitly references "a subset of DHCPv6" and cites RFC 3736 for the definition of that subset. However, as Greg points out, it's not necessarily the case that RFC 3736 was used in the implementation of that subset.
- Ralph
At 12:46 PM 8/12/2004 +1000, Greg Daley wrote:
Hi Daniel,
S. Daniel Park wrote:=> Right, but there is no need to have the O flag off. To me RFC 3736 is something useful for server vendors and should not be associated with
setting the O flag.
You mean we can always set O flag ? I don't make sense why RFC3736 should not be associated with setting the O flag.
It does make sense, since RFC 3736 is for servers and relays, not hosts.
If an RFC 3315 server is available on a network, but network policy discourages use of stateful addresses, we may wish to advertise O=on, M=off.
RFC2461bis says O flag (when set) indicates a subset of DHCPv6 [RFC3736] is available for autoconfiguration of other
(non-address) information...
This is wrong.
a subset of DHCP is available, but this doesn't imply RFC3736.
Greg
-------------------------------------------------------------------- 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 --------------------------------------------------------------------
