IETF-SSH archive

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

Re: Traffic Analysis



Markus Friedl wrote:
On Mon, Jan 17, 2005 at 12:14:40AM +0000, Ben Harris wrote:

In article <Pine.HPX.4.58.0501161506080.12089%edison.cisco.com@localhost> you write:

9.2.9  Traffic Analysis

 Passive monitoring of any protocol may give an attacker some
 information about the session, the user, or protocol specific
 information that they would otherwise not be able to garner.  For
 example, it has been shown that traffic analysis of an SSH session
 can yield information about the length of the password.  [Openwall]
 Implementors should use the SSH_MSG_IGNORE packet as described in
 [SSH-TRANS] along with any other methods they may find to prevent
 traffic analysis.

It might also be worth mentioning that the "random padding" field can be
used to obscure the length of packets.  I'd suggest, after "SSH_MSG_IGNORE
packet" adding ", and variable-length random padding,", though that wording
could doubtless be improved.


please don't add this, random padding is wrong and
does not really help, as it can be filtered out.

I'm not sure I understand this?  How can it be filtered out?
I thought this attack was based on the fact that in most
implementations, for certain key packets, the TCP segment
length (which is observable) often corresponds to the the
SSH packet length (which is not observable), and therefor
allows certain deductions about the length of sensitive
fields with-in the packet?

If that is the case, then it seems to me that perturbing the
pad to change the length of the the packet (and therefore
the length of the TCP segment) would remove the attackers
ability to make such deductions about the length of fields
in the packet?

It also seems that it is more likely that a packet with extra
pad will be sent as one TCP segment than that a
SSH_MSG_USERAUTH_REQUEST and a SSH_MSG_IGNORE will be sent
in a single TCP segment (and probably easier to implement.)

In fact, I believe the wording above needs some adjustment
to note that for SSH_MSG_IGNORE to be effectively used
to thwart this kind of attack, it MUST be sent in the same
TCP segment as the SSH packet it is protecting.

Thanks,

Joseph



Home | Main Index | Thread Index | Old Index