RE: Caching a sniffer

["Shawn Jackson" <sjackson () horizonusa com>]

> > In essence if you flood the MAC table of a switch the switch will
turn 
> > into a hub, thus "disabling the switch component of the ports".

> Of course, that's not necessarily true.  The behavior of a switch when
the MAC 
> address table is exceeded is not defined by any standard, nor is it
often specified 
> by the manufacturer.

Of course your correct. But I've tested numerous switches, Cisco being
the most
tested. Bay Networks, SMC, HP/Compaq network gear have all responded in
a somewhat
similar fashion.

> 1. Dump the entire MAC table.  Switch acts as if power on reset just
occurred.

Seams logical, but I've never seen it implemented. It would halt traffic
while
learning resumes, in addition if other checks needed to run (Spanning
Tree/CDP)
it would take much longer.

> 2. Stop learning.  All previously learned MAC addresses remain, and so
only traffic 
> for unrecognized MAC addresses gets sent to all ports.

That would damage the network. If a new client fires up, they would not
get added to
the switches tables and not receive any traffic.

> 3. Partial Purge of table.  Some portion of the table gets purged and
the switch 
> continues, treating those purged MAC addresses as if this was the first
time they 
> were seen.  Depending upon how the purged addresses are selected -
oldest first, 
> youngest first, random, lowest MAC addresses, highest MAC addresses or
something 
> else - will cause the switch to act differently for different users.

Seams a better solution out of the bunch, could be a pain to implement.

> 4. Shutdown port - assume hostile intent and stop forwarding traffic.

Called Port-Security on Cisco switches, don't know about the others.

> Do I know of which switches do what?  Nope.  But we should ALL have
learned the 
> lessons of depending upon undocumented behaviors and unspecified
conditions with Y2K.

Agreed, but all we need to do is ask the manufacturer of the equipment
to see
how it deals with the problems/attacks we are describing. It's not, "the
way it works"
it's how they dealt with the problem. Cisco's do it by design, and so do
most manufactures.

> SPAN/Port Mirroring/Roving Analysis Port(3Com) is intentional and
controlled by the 
> administrator.  Also, how the port handles traffic in excess of it's
capacity 
> (say you are monitoring 3 100BaseT ports out a single 100BaseT port),
is completely 
> Mfg dependent and undocumented.

Traffic gets queued, just like any other switch-port. FiFo, Weighted,
etc. That's not
undocumented, it's just not advertised on most switches. On the better
ones you can
select your queuing strategy.

Shawn Jackson
Systems Administrator
Horizon USA
1190 Trademark Dr #107
Reno NV 89521

www.horizonusa.com
Email: sjackson () horizonusa com
Phone: (775) 858-2338
       (800) 325-1199 x338

---------------------------------------------------------------------------
Ethical Hacking at the InfoSec Institute. Mention this ad and get $545 off
any course! All of our class sizes are guaranteed to be 10 students or less
to facilitate one-on-one interaction with one of our expert instructors.
Attend a course taught by an expert instructor with years of in-the-field
pen testing experience in our state of the art hacking lab. Master the skills
of an Ethical Hacker to better assess the security of your organization.
Visit us at:
http://www.infosecinstitute.com/courses/ethical_hacking_training.html
----------------------------------------------------------------------------