Date: Thu, 31 May 2001 17:34:47 +0900
From: JINMEI Tatuya <[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
| Here are some inconsistency comparing to the latest versions of the
| scoping/address architecture drafts:
Here I think it is better to use the values from the currently published
RFCs - the actual numbers aren't important to anything in this draft,
and using data from a doc that isn't yet an RFC will only serve to hold
up this one until that is eventually published.
Just use what exists now, and then allow the later doc to update the
numbers when it appears.
| I think "no route to DB" here is a bit confusing for IPv6 hosts, based
| on the neighbor discovery specification. Section 5.2 of RFC 2461
| says:
That section can really only apply after interface selection (routing)
has been done - that is, it is assuming that a link has been selected
to send the packet upon, at that stage one can compare prefixes, select
the next hop, or send it as on-link traffic. But before that is done,
something has to select the interface. Even for hosts with only one
piece of electrical gizmo called an interface (or NIC) is likely to have
at least an internal software "loopback" interface, and perhaps some
tunnels as well. Something has to have selected one of those, and during
that selection there's always the possibility that "none of the above"
is the answer obtained. I suspect that's what the "no route to DB"
is referring to (that's certainly the way I understood it). For many
hosts that answer will never occur, and that section of the address
selection draft won't be helpful - for others, it can occur.
kre
--------------------------------------------------------------------
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]
--------------------------------------------------------------------
