And the reason (which Dave neglected to mention) is that Section 6.3.7 of
RFC 2461 and 2461bis specifies precisely five conditions that allow a host
to send an RS and forbids the host from sending an RS at any other time.
Expiration of the RA lifetime is not one of those conditions.
jak
----- Original Message -----
From: "Dave Thaler" <[EMAIL PROTECTED]>
To: "Alexandru Petrescu" <[EMAIL PROTECTED]>; "Narayanan,
Vidya" <[EMAIL PROTECTED]>
Cc: "Templin, Fred L" <[EMAIL PROTECTED]>; "James Kempf"
<[EMAIL PROTECTED]>; "INT Area" <[EMAIL PROTECTED]>
Sent: Wednesday, August 09, 2006 9:51 AM
Subject: RE: [Int-area] IPv6 addressing model, per-MN subnet prefix,and
broadcast domain
Alexandru Petrescu writes:
> I'm not exactly sure what you mean by RAs with "incomplete
> information". In any case, the above design seems weird to me. For
> one, there is no reason a host should send an RS right after it
> received an unsolicited RA.
Well, if the router puts a short lifetime in the periodic partial RA
(partial, incomplete would mean only the default router address) then
the host will necessarily do an RS after that lifetime and get an RA
that would contain only prefix (no default router address) with a much
longer lifetime. I think it's implementable.
Only if you modify the hosts to introduce new behavior, as what you
describe above is not the RFC 2461 specified behavior.
-Dave
_______________________________________________
Int-area mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/int-area
