Re: Errata for RFC4862

[神明達哉 <[e-mail]>]

At Sun, 8 Jan 2017 16:30:04 +1100,
Mark Smith <[e-mail]> wrote:

> An attacker could take advantage of this by doing the following:
>
> 1. Send an initial RA with a ValidLifetime of 2 hours, which would pass Check 1.
> 2. Wait until around 1 hour and 59.5 minutes, and then send an RA with
> a ValidLifetime of 61 seconds.

I guess the unstated assumption behind this defense of RFC4862 is that
at least some legitimate RA containing a PIO with a sufficiently long
valid lifetime updates the lifetime during this 1-hour-and-59.5-minute
period.  Of course, we cannot always assume this, but as I believe
everyone understands, this defense is supposed to be quite basic and
weak anyway, and does not intend to prevent all kinds of DoS attempt
by most sophisticated attackers under all possible circumstances.

In that sense the current text of RFC4862 actually does what it
intended to do with originally perceived limitations, so I don't see
the need for the proposed update.

At the very least I don't think this is an editorial error.  If we now
agree that the currently described defense is too weak, it should be
addressed as a protocol update to the RFC.

--
JINMEI, Tatuya