Re: ssh and ids

[Adam Powers <apowers () lancope com>]

There is really no one full-proof answer to this question (that I'm aware
of). Encryption remains the bane of network-based intrusion detection
technologies.

At the risk of speaking on behalf of such flow-based vendors as Arbor, Mazu,
Q1, and (yes, my personal favorite) Lancope, I think some of the new
behavioral traffic analysis technologies go a long way toward solving some
of the problems presented by encryption technologies.

<light details>
By observing the duration of a "flow" (read: a TCP socket or series of
related sockets) and the manner in which packets are exchanged over a "long
duration" flow, a behavior-based system can pinpoint those connections that
seem to be "out of the norm". During the baselining period, a behavior
driven system observes connections attributes such as "duration" and
"relative connectedness" to gain an understanding of the nature of the flows
being created by a given network node. The flow-based, behavior-driven
system should have the ability to discern between a AES gotomypc.com
connection over TCP 443 and an automatic refresh connection to
www.weather.com. The determination that "covert communications" are underway
is done not through string matching or protocol anomaly but rather through
the analysis of the flow attributes themselves (duration, packets sent/rcvd,
pkt size, etc). Bottoms line: the magic is in the algorithms used to examine
header traffic. Header traffic is not encrypted.
</light details>

The #1 defining attribute of flow-analysis techniques is that they typically
DO NOT require use of payload data to determine the presence of an attack.

As previously mentioned, there is no fool-proof plan... Flow-based
technologies can be tricked... It just requires a much different science
than that used by snot, sidestep, or encrypted shell shoveling.

- AP



On 6/18/04 2:18 PM, "Runion Mark A FGA DOIM WEBMASTER(ctr)"
<mark.runion () us army mil> wrote:

> Lets suppose the attacker is mildly sophisticated, and after making the
> initial assault roots the box and installs a secure backdoor or two.  Is
> there any IDS capable of isolating data it cannot read, except to monitor
> authorized port usage of a system or group of systems?  Not to complicate
> the question, but when the attacker is using portal gates and all
> communications traffic is encrypted in normal channels how can an IDS
> participate?  Monitoring normal traffic patterns seems a bit slow for
> detection.
>
> -
> Mark Runion 
>
>
> ---------------------------------------------------------------------------
>
> ---------------------------------------------------------------------------



---------------------------------------------------------------------------

---------------------------------------------------------------------------