At Wed, 2 Oct 2002 16:08:34 -0700, Steve Deering wrote:
>
> At 6:07 PM -0400 10/2/02, Rob Austein wrote:
> >The key phrase in your explanation is "the admin assigns". The
> >addr-arch doc says "admin-local scope is the smallest scope that must
> >be administratively configured". So which is it?
>
> You omitted the full description:
>
> admin-local scope is the smallest scope that must be
> administratively configured, i.e., not automatically
> derived from physical connectivity or other, non-
> multicast-related configuration.
>
> Subnet-local scope is an example of automatic derivation from "other,
> non-multicast-related configuration".
^
administrative
> Specifically, you don't directly configure a router to know which
> subnet-scope boundaries pass through it (as you must do with larger
> scopes). Rather, you (typically, manually) configure the router
> with subnet info -- including, perhaps, enabling or disabling
> multilink-subnet behavior -- as required for unicast routing, and
> then you automatically derive subnet-scope boundary information from
> that "other, non-multicast-related configuration".
>
> Or saying it more concisely: you don't administratively configure
> subnet scope; you adminstratively configure subnet info for unicast
> purposes, and then automatically derive subnet scope from that.
Fine, you don't do multicast configuration, you do non-muliticast
configuration and automatically derive the multicast configuration of
it from the non-multicast configuration.
The point, however, if you go back to my original message, is that the
text in the draft can also be read as saying that the router is
somehow supposed to deduce this bit of configuration from its physical
connectivity without any administrative configuration at all, which is
sufficiently nebulous that it could logic like "oh, I have an ethernet
interface and a firewire interface, and I just know that multi-link
subnets were invented to make my firewire interface happy, so set the
same scope ID on both of these interfaces and forward packets between
them." Which (again suspending disbelief) would arguably be ok if
there were a specification for multi-link behavior that said to do
that, but there isn't, so the draft leaves some ambiguity about
whether certain packets should be forwarded or not.
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
