>>>>> On Wed, 04 Feb 2004 10:17:44 +0100,
>>>>> Francis Dupont <[EMAIL PROTECTED]> said:
> The current RFC2462 describes in Section 5.5.3 e) how the valid
> lifetime of an autoconfigured address is updated, considering the
> avoidance of DoS attack with too short lifetimes.
> => the DoS attack is about valid lifetime only because when a valid
> lifetime is dropped to zero an implementation can (IMHO it is in fact
> a SHOULD) kill all connections using a now unvalid address.
> The behavior for zero preferred lifetime is far less drastic ("deprecated"
> addresses are not used for new "connections" when there are other
> available addresses) and (IMPORTANT) we need for multi-homing to be
> able to play with preferred lifetimes.
> However, it doesn't mention preferred lifetimes. 5.5.3 e) says:
> ...
> This document doesn't say anything about preferred lifetimes from this
> part to the end of this section.
> => this is on purpose.
I know the rationale why the two-hour rule only affects the valid
lifetime.
Perhaps I was not clear enough in the previous message, but my point
is that RFC2462 doesn't say **anything** about preferred lifetime
update.
To make the story simple, let's forget the DoS issue and the two-hour
rule for now. Assume we receive a valid RA containing a prefix
information option, and the processing the RA reaches at 5.5.3 e).
Also assume the valid and preferred lifetimes of the prefix
information are long enough so that I don't have to worry about a DoS.
Now please re-read 5.5.3 e) carefully. It is very clear that we
should reset the valid lifetime to the advertised valid lifetime. But
we cannot be sure how we should do about the preferred lifetime
because 5.5.3 e) doesn't say anything about the preferred lifetime.
Of course, a common sense would suggest that we also reset the
preferred lifetime to the advertised value, and I believe all
implementations work this way, at least in normal (i.e., non-DoS)
cases. Still, it's just a guess; RFC2462 doesn't say anything about
this point.
So my first point is that we should clearly specify how the preferred
lifetime is updated in 5.5.3 e) of rfc2462bis, mainly for normal
cases. My second point is what we should do about the preferred
lifetime when the valid lifetime is ignored due to the two-hour rule.
My suggestion to the first point is that the preferred lifetime should
basically be reset to the advertised value (of course!).
My suggestion to the second point is that we should also ignore the
preferred lifetime if the valid lifetime is ignored due to the
two-hour rule.
Hope I'm clear this time.
JINMEI, Tatuya
Communication Platform Lab.
Corporate R&D Center, Toshiba Corp.
[EMAIL PROTECTED]
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[EMAIL PROTECTED]
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
