DEVS Simulation of distributed intrusion detection systems

  • Authors:
  • Tae Ho Cho;Hyung Jong Kim

  • Affiliations:
  • School of Electrical Computer Engineering, SungKyunKwan University, Suwon, Korea;Korea Information Security Agency, Seoul, Korea

  • Venue:
  • Transactions of the Society for Computer Simulation International - Recent advances in DEVS Methodology--part I
  • Year:
  • 2001

Quantified Score

Hi-index 0.00

Visualization

Abstract

An intrusion detection system (IDS) attempts to identify unauthorized use, misuse, and abuse of computer and network systems. As intrusions become more sophisticated, dealing with them moves beyond the scope of one IDS. The need arises for systems to cooperate with one another, to manage diverse attacks across networks. The feature of recent attacks is that the packet delivery is moderately slow, and the attack sources and attack targets are distributed. These attacks are called "stealthy attacks." To detect these attacks, the deployment of distributed IDSs is needed. In such an environment, the ability of an IDS to share advanced information about these attacks is especially important. In this research, the IDS model exploits blacklist facts to detect the attacks that are based on either slow or highly distributed packets. To maintain the valid blacklist facts in the knowledge base of each IDS, the model should communicate with the other IDSs. When attack level goes beyond the interaction threshold, ID agents send interaction messages to ID agents in other hosts. Each agent model is developed as an interruptible atomic-expert model in which the expert system is embedded as a model component.