Hi
Adam,
Very good point, I knew I was forgetting this point. I will try to add some test to the update about this.
I think that that stateful address config is probably conditionally
mandatory (or even unconditionally
mandatory. The section on 2462 needs this info. Also, we need
some extra text in DHCPv6 to discuss this,
as
well.
Thanks,
John
-----Original Message-----
From: ext Adam Machalek [mailto:[EMAIL PROTECTED]]
Sent: 21 June, 2002 18:52
To: [EMAIL PROTECTED]
Subject: Stateful Address Config - I-D ACTION:draft-ietf-ipv6-node-requirements-00.txt
John,
Section 5.4 on DHCPv6 states that DHCPv6 is unconditionally optional, which today implicitly makes Stateful Address config optional.
However, I'd like to see some direct clarification in this document on Stateful Address config, probably directly after Section 4.5.2 discussing Stateless Address config.
RFC2462 has some ambiguities, in particular, it states "a managed address configuration flag indicates whether hosts should use stateful autoconfiguration", not SHOULD. Later in 2462 in section 5.2 it continues this ambiguity by never explicitly using MAY/SHOULD/MUST anywhere.
Still later in section RFC2462 5.5.2 Absence of Router Advertisements, it finally states that in the absence of a router, a host MUST attempt to use stateful autoconfiguration.
And lastly, the DHCPv6 draft states "DHCP is one vehicle to perform stateful autoconfiguration", implying that there may be others.
So in the end, even with DHCPv6 optional, this doesn't clarify exactly what a host should do about Stateful Address autoconfiguration.
Adam Machalek
