[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Evaluation: draft-ietf-mobileip-ipv6 - Mobility Support in IPv6to Proposed Standard



--On torsdag, mars 06, 2003 10:31:19 -0500 "Steven M. Bellovin" <smb@research.att.com> wrote:

In message <540320000.1046964395@askvoll.hjemme.alvestrand.no>, Harald
Tveit Alvestrand writes:


--On torsdag, mars 06, 2003 00:17:34 -0500 "Steven M. Bellovin"
<smb@research.att.com> wrote:


9.5.1   If there's no IPsec-level replay protection, this sequence
number just won't do the trick.  A wireless mobile node could very
        easily generate enough binding updates per day that an enemy
        could replay old ones that appeared to be in the window.

32767 binding updates in a day?
that's 22 moves per minute, or one every 3 seconds, sustained.




I didn't say "per day".  But what about 6 months of driving with a
MobileIP-based PDA or cell phone?  That's one cell site every 6.5
minutes, which is not unreasonable.
you had "per day" in there, which was what confused me.
So some kind of timestamp would also protect against replay attacks.
Not because I have understood this, but the section also says:

If the receiving node no longer recognizes the Home Nonce Index
value, Care-of Nonce Index value, or both values from the Binding
Update, then the receiving node MUST send back a Binding
Acknowledgement with status code 136, 137, or 138, respectively.

The MAX_NONCE_LIFETIME is 240 seconds.
So this may not be a problem, but seems to confirm my impression that Mobile IPv6 is a VERY chatty protocol....

anyway, the Mobile IPv6 experts should be able to answer one way or the other whether this is a problem.

Harald