Arquitetura multi-agentes para detecção de intrusão distribuida

Abstract

The growing concern about information security in computer networks is responsible for constantly producing new ways to defend them. Within this context, the development of new ways of intrusion detection plays an important role in protecting the information. Detection systems must be efficient and, at the same time, must not overload the network or the processing capabilities of the nodes within it. In order to be effective, a system must base its decisions on as many sources of information as possible and organize knowledge in a way that allows a functional communication between those sources. This dissertation describes the proposal for a Distributed Intrusion Detection System architecture that uses mobile agents and an ontology for information sharing. Mobile agents provide a convenient way to distribute the detection process, enabling peer to peer cooperation between network nodes without generating much additional traffic. The ontology provides an organized way of storing and sharing knowledge. The proposed architecture has been implemented using the Java programming language and JADE framework and a test laboratory has been assembled to verify the operation of the system. The tests results confirmed that a distributed multi-agent architecture that uses an ontology can be effective in detecting attacks on networks and systems.