Re: DHCP vs. RA... again.

[Jari Arkko <jari.arkko@piuha.net>]

Brian E Carpenter kirjoitti:
>> ... Of course, this means that a good case has
>> to be made.
>
> It was made a long time ago I think, since we always
> expected strongly managed IPv6 networks where DHCP
> is more suitable, and loosely managed networks
> (the proverbial dentist's office network) where it
> isn't. Obviously there's a grey area in the middle
> where either approach would work. What's changed in
> this argument?

Not necessarily anything -- I'm not arguing against the
case, I'm arguing that we need a good justification. In the
beginning of thread, there seemed to be a jump from
IETF-69 network experiences to something quite
different, and IMHO that was not a sufficient
justification :-)

Anyway, the managed/unmanaged networks certainly are real
business situations.

Ralph and DHC are working on what the possible
case for DHCP-side prefix/default router information
would be. Some of the questions I would like to
see answered include whether there's a true
change in vulnerability to accidental/spoofed
RAs, as similar issues may also happen in DHCP.

For the DNS etc. information delivery in RAs: do we
have (a) proposals for going beyond draft-jeong,
other than changing its status, and (b) rationale why
the status upgrade or the extensions are needed?
I'm hoping the latter would also explain what we
expect to get that goes further than what
stateless-DHCP-implemented-in-a-router would
provide.

Jari