DARPA is funding an effort to combine state of the art intrusion detection systems with intelligent firewalls and routers to form an intranet wide automated defense system. The basic idea is to set these components up in a virtual security network where security components talk to their neighbors. The virtual security network has some correspondence with the real network topology because the firewalls and routers separate the real network into different regions.
Using this system, a variety of attacks will be detected. Upon an attack being launched, security components will alert each other of the attack and a component will be selected to initiate an automated response. Automated responses will vary with the attack but so far involve some sort of filtering of network packets.
The system described so far has been implemented and tested. Three different corporate entities have been brought together to produce the defense system: Boeing corporation, Trusted Information Systems, and UC Davis. Boeing has supplied the intelligent routers as well as the intruder detection and isolation protocol (IDIP) which enables the components to communicate. Trusted information systems has supplied a modified version of their firewall toolkit to work in this environment. And UC Davis has provided the intrusion detection system called the master intrusion detection system (MIDS).
The first contract was to build and test the system described. We have two new contracts to continue this work. The first contract will be to explore the idea of automated response and figure out what is an intelligent way to respond to reports of attacks. After all, we don't want the response to be more harmful than the attack itself. The second contract will be to harden this system, integrate it into other projects, and create a production quality defense system. This final system will be installed at various air force bases.
Some Deliverables: Response engine Graphical user interface Aggregation engine Host-based response tookit Attack description language Abstract model of system
Contact person: Ricardo Anguianoanguiano@cs.ucdavis.edu