One (perhaps last) question about the M/O flags for rfc2462bis:

Currently, RFC2462 uses the term "stateful" as the counter part of
the "stateless" configuration defined in RFC2462, both for address
configuration (the M flag) and for other configuration (the O flag).

Using "stateful" should be okay for address configuration (the M flag
part).

However, as Ralph pointed out before, "stateful" may not be
appropriate for other configuration information:
https://www1.ietf.org/mail-archive/working-groups/ipv6/current/msg02262.html

In particular, the fact that RFC3736 (which we are primarily
considering as the protocol for the O flag) is entitled "*Stateless*
Dynamic Host Configuration Protocol (DHCP) Service for IPv6" will
confuse implementors if we keep calling it "stateful" in rfc2462bis.

So, I strongly believe we should clarify this point in some way.
Possible solutions would include:

1. remove "stateful" from the definition of the O flag (in
   rfc2461bis), that is, change

      O              1-bit "Other stateful configuration" flag.  When
                     set, hosts use the administered (stateful) protocol
                     for autoconfiguration of other (non-address)
                     information.  The use of this flag is described in
                     [ADDRCONF].
      (RFC2461 Section 4.2)

   to (e.g.):

      O              1-bit "Other configuration" flag.  When
                     set, hosts use a separate protocol
                     for autoconfiguration of other (non-address)
                     information.  The use of this flag is described in
                     [ADDRCONF].

   and reword rfc2462bis accordingly.

2. do not touch the definition of the O flag, but add notes for
   clarification in rfc2462bis like this:

      While the flag and the corresponding protocol are called
      "stateful" in order to highlight the contrast to the stateless
      protocol defined in this document, the intended protocol
      [RFC3736] is also defined to work in a stateless fashion.  This
      is based on a result, through experiments, that all known
      "other" configuration information can be managed by a stateless
      server, that is, a server that does not maintain state of each
      client that the server provides with the configuration
      information.

I personally prefer the former with small preference since it should
be a cleaner clarification.  But I can live with the second approach,
too.

What do others think?  Is there any other opinions?

                                        JINMEI, Tatuya
                                        Communication Platform Lab.
                                        Corporate R&D Center, Toshiba Corp.
                                        [EMAIL PROTECTED]

--------------------------------------------------------------------
IETF IPv6 working group mailing list
[EMAIL PROTECTED]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------

Reply via email to