Date: Fri, 02 Aug 2002 10:40:18 +0200
From: Brian E Carpenter <[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
| I think in the end this is an argument about nothing.
I think I agree with that.
| > This implies that IPv6 nodes SHOULD NOT establish any flow-specific
| > state unless so instructed by a specific flow state establishment method.
|
| What Margaret and itojun refer to are flow state establishment methods that
| happen to consist of algorithms built into the sending node (rather than
| signaling mechanisms or pre-configured flow states). These algorithms
| must avoid duplicate labels, just as signaled or pre-configured
| methods must.
Yes, when you consider only the source nodes. There the method you
describe is correct, and the concerns expressed aren't a problem.
The wording could perhaps be improved to make that clear.
But...
| So the sentence is in fact a no-op and can be deleted.
I don't agree with that - I read the relevant sentence as being directed
at intermediate nodes (routers), not at the end nodes (hosts). That is,
I thought that it was trying to say that routers should not just "notice"
that there seems to be a flow passing by, and attempt to do magic things
to that flow, because there's no way to flush the state they establish to
do that. That is, there needs to be some kind of setup mechanism (anything
from RSVP to an RFC) that defines just how the flows are established, and
when they should be torn down again.
Don't delete it, just make its scope clearer.
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]
--------------------------------------------------------------------
