RE: Use of Taps for IDS
From: Jackie Chan (blue0ne@bello.digitz.org)Date: 03/13/02
- Previous message: Joshua Krage: "Re: Use of Taps for IDS"
- In reply to: Hector Herrera: "RE: Use of Taps for IDS"
- Next in thread: Bob Walder: "RE: Use of Taps for IDS"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Date: Tue, 12 Mar 2002 21:38:52 -0500 (EST) From: Jackie Chan <blue0ne@bello.digitz.org> To: Hector Herrera <hectorh@pobox.com>
The IDS/TAP config can be improperly set up , just like any other system.
But if you follow the instructions in the following paper (written circa
1999) you will not get anything near what you are describing.
http://www.digitz.org/IDS/Shomiti.pdf
-blue0ne
On Tue, 12 Mar 2002, Hector Herrera wrote:
> At 10:45 AM 12/03/02 +0000, Bob Walder wrote:
> >Can I just throw in another little "gotcha" here too? It's not really about
> >taps, but port mirroring, so I guess it's sorta on topic...
> >
> >Imagine you have your 100/1000Mbit switch mirroring ports 1-10 to port 11,
> >where your IDS is. So far so good, and any attack traffic coming on to the
> >switch from elsewhere, targeted at the mirrored ports, will be picked up by
> >the IDS.
> >
> >But what happens when port 1 launches some sort of attack against port 2?
> >The infamous "disgruntled employee" strikes again.
> >
> >Now we see TWO alerts for every attack - one coming FROM port one and one
> >coming TO port 2. A simple Sniffer trace on port 11 of a Ping from port 1 to
> >port 2 reveals:
> >
> >ICMP echo request from 1 to 2 (as seen mirrored from port 1)
> >ICMP echo request from 1 to 2 (as seen mirrored from port 2)
> >ICMP echo reply from 2 to 1 (as seen mirrored from port 2)
> >ICMP echo reply from 2 to 1 (as seen mirrored from port 1)
> >
> >Two distinct streams, since a SPAN port does not "normalise" traffic from a
> >single session (are there switches that can do this?). What to do with these
> >duplicate requests?
>
> I use a cisco 2912-XL with port-mirroring for
> my snort logger, and I have not seen the behaviour
> that you describe.
>
> I believe that is because the switch that I have
> copies the packet from the switching fabric, not
> from the input/output queues.
>
>
>
- Previous message: Joshua Krage: "Re: Use of Taps for IDS"
- In reply to: Hector Herrera: "RE: Use of Taps for IDS"
- Next in thread: Bob Walder: "RE: Use of Taps for IDS"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Relevant Pages
|