A modular architecture for distributed IDS in MANET

  • Authors:
  • Ricardo S. Puttini;Jean-Marc Percher;Ludovic Mé;Olivier Camp;Rafael De Sousa;Cláudia J. Barenco Abbas;L. Javier García-Villalba

  • Affiliations:
  • Department of Electric Engineering, University of Brasilia, Brasilia, DF, Brazil;École Supérieure d'Électronique de l'Ouest, France, Angers Cedex 01, France;Supélec, Cesson Sévigné Cedex, France;École Supérieure d'Électronique de l'Ouest, France, Angers Cedex 01, France;Department of Electric Engineering, University of Brasilia, Brasilia, DF, Brazil;Department of Electric Engineering, University of Brasilia, Brasilia, DF, Brazil;Department of Computer Systems and Programming, Complutense University of Madrid, Madrid, Spain

  • Venue:
  • ICCSA'03 Proceedings of the 2003 international conference on Computational science and its applications: PartIII
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper we propose a distributed and modular architecture for an intrusion detection system (IDS) dedicated to a mobile ad hoc network (MANET) environment. The main feature of our proposition relies on the use, on each node of the MANET, of a local IDS (LIDS) cooperating with other LIDSes through the use of mobile agents. The modular design is needed as a response to the extensibility requirements related to the complex contexts of MANET. The proposed solution has been validated by a proof-of-concept prototype, which is described in the paper. Two different types of attacks are presented and have been implemented, at the network level and at the application level. The detection of such attacks are formally described by specification of data collection, attack signatures associated with such data and alerts generation, emphasizing the relation of each of these detection steps with the modules in the designed architecture. The use of the management information base (MIB) as a primary data source for the detection process is discussed and modules for MIB data extraction and processing are specified and implemented in the prototype. Experiments exhibit fairly good results, the attacks being collaboratively detected in real-time.