A Distributed Intrusion Detection System Using Cooperative Agents

Due to the extension of attacks to computer networks in the current decade, it is crucial to secure the networks in an appropriate and effective way. One solution is to have a host that supervises the situation for all the computers in the network and makes decision regarding possible attacks. This method is not effective in recently developed networks due to their large extension and high speed. We propose a new distributed intrusion detection system that is based on cooperative agents (‘DIDCA’). Agents are grouped into subsets of agents that are in connection with a leader on top. Leaders of different subsets communicate together to share all the information throughout the network. We simulated our method using DARPA 1999 database and compared it with Snort. Our results showed that our method achieved a better performance.