Vulnerability Development mailing list archives

Re: DHCP and Security


From: sebastion () IRELANDMAIL COM (Jeff Bachtel)
Date: Sat, 5 Feb 2000 12:44:44 -0600


Hmm, at the .edu I work for (which switched to DHCP about a year ago),
the DHCP server hands out to new MAC addresses a non-routable ip
address (such as 10.0.0.1 or whatever). There's a huge pool of those
addresses (probably with near-non-existent lease times), which are
absolutely useless, that is the only thing you can do with such a
source ip is connect to the DHCP server (which also acts as a web
server), and authenticate yourself (via ph). Ones you are
authenticated, the DHCP server "knows" your MAC address (actually, its
put into a SQL database) and will upon your next attempt to grab an
address, give you a valid, routable ip.

Most of the stuff is custom, but shouldn't take a competent coder more
than a few weeks to code and test out (assuming you already have some
sort of database [LDAP/SQL/etc] to auth against), and eliminates the
ability to deplete the pool of routable ip addresses.

jeff

On Thu, Feb 03, 2000 at 01:16:08PM -0500, Nitzenberger, Rob, MSgt, AF/XORR wrote:
Need a policy read folks:

The system I "manage" has 3200 users at various locations throughout the
world, managed by a central NOC.  Our firewall permissions (protocol and
port) are highly restrictive and report any unauthorized actions (ftp,
pings, finger,.....).  The NOC gets a report from the firewall indicating
which IP was the "offender".  If I the LAN clients are configured with
static IP's, it's easy to attribute the offending action with a LAN client,
but with DHCP (which is the method of choice for our sys admin types), it
has proven difficult to "map" an IP address back to a specific user... lease
times expire, inadequate event logging..etc.

 How can I configure DHCP to balance the need for security with the wishes
of the sys admin folks?  Any Ideas?

Rob Nitzenberger
thenitz () email com


Current thread: